notebooks/L03_pandas/v02_pandas_basics.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Basic Functionality\n",
    "\n",
    "**Prerequisites**\n",
    "\n",
    "- [pandas Intro](intro.ipynb)  \n",
    "\n",
    "\n",
    "**Outcomes**\n",
    "\n",
    "- Be familiar with `datetime`  \n",
    "- Use built-in aggregation functions and be able to create your own and\n",
    "  apply them using `agg`  \n",
    "- Use built-in Series transformation functions and be able to create your\n",
    "  own and apply them using `apply`  \n",
    "- Use built-in scalar transformation functions and be able to create your\n",
    "  own and apply them using `applymap`  \n",
    "- Be able to select subsets of the DataFrame using boolean selection  \n",
    "- Know what the “want operator” is and how to apply it  \n",
    "\n",
    "\n",
    "**Data**\n",
    "\n",
    "- US state unemployment data from Bureau of Labor Statistics  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "## Outline\n",
    "\n",
    "- [Basic Functionality](#Basic-Functionality)  \n",
    "  - [State Unemployment Data](#State-Unemployment-Data)  \n",
    "  - [Dates in pandas](#Dates-in-pandas)  \n",
    "  - [DataFrame Aggregations](#DataFrame-Aggregations)  \n",
    "  - [Transforms](#Transforms)  \n",
    "  - [Boolean Selection](#Boolean-Selection)  \n",
    "  - [Exercises](#Exercises)  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "## State Unemployment Data\n",
    "\n",
    "In this lecture, we will use unemployment data by state at a monthly\n",
    "frequency."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'0.25.1'"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "%matplotlib inline\n",
    "# activate plot theme\n",
    "import qeds\n",
    "qeds.themes.mpl_style();\n",
    "\n",
    "pd.__version__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "First, we will download the data directly from a url and read it into a pandas DataFrame."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "## Load up the data -- this will take a couple seconds\n",
    "url = \"https://datascience.quantecon.org/assets/data/state_unemployment.csv\"\n",
    "unemp_raw = pd.read_csv(url, parse_dates=[\"Date\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "The pandas `read_csv` will determine most datatypes of the underlying columns.  The\n",
    "exception here is that we need to give pandas a hint so it can load up the `Date` column as a Python datetime type: the `parse_dates=[\"Date\"]`.\n",
    "\n",
    "We can see the basic structure of the downloaded data by getting the first 5 rows, which directly matches\n",
    "the underlying CSV file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Date</th>\n",
       "      <th>state</th>\n",
       "      <th>LaborForce</th>\n",
       "      <th>UnemploymentRate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>Alabama</td>\n",
       "      <td>2142945.0</td>\n",
       "      <td>4.7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>Alaska</td>\n",
       "      <td>319059.0</td>\n",
       "      <td>6.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>Arizona</td>\n",
       "      <td>2499980.0</td>\n",
       "      <td>4.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>Arkansas</td>\n",
       "      <td>1264619.0</td>\n",
       "      <td>4.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>California</td>\n",
       "      <td>16680246.0</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Date       state  LaborForce  UnemploymentRate\n",
       "0 2000-01-01     Alabama   2142945.0               4.7\n",
       "1 2000-01-01      Alaska    319059.0               6.3\n",
       "2 2000-01-01     Arizona   2499980.0               4.1\n",
       "3 2000-01-01    Arkansas   1264619.0               4.4\n",
       "4 2000-01-01  California  16680246.0               5.0"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp_raw.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "Note that a row has a date, state, labor force size, and unemployment rate.\n",
    "\n",
    "For our analysis, we want to look at the unemployment rate across different states over time, which\n",
    "requires a transformation of the data similar to an Excel pivot-table."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Alabama</th>\n",
       "      <th>Alaska</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>Arkansas</th>\n",
       "      <th>California</th>\n",
       "      <th>Colorado</th>\n",
       "      <th>Connecticut</th>\n",
       "      <th>Delaware</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Georgia</th>\n",
       "      <th>...</th>\n",
       "      <th>South Dakota</th>\n",
       "      <th>Tennessee</th>\n",
       "      <th>Texas</th>\n",
       "      <th>Utah</th>\n",
       "      <th>Vermont</th>\n",
       "      <th>Virginia</th>\n",
       "      <th>Washington</th>\n",
       "      <th>West Virginia</th>\n",
       "      <th>Wisconsin</th>\n",
       "      <th>Wyoming</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>4.7</td>\n",
       "      <td>6.3</td>\n",
       "      <td>4.1</td>\n",
       "      <td>4.4</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.8</td>\n",
       "      <td>2.8</td>\n",
       "      <td>3.5</td>\n",
       "      <td>3.7</td>\n",
       "      <td>3.7</td>\n",
       "      <td>...</td>\n",
       "      <td>2.4</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.6</td>\n",
       "      <td>3.1</td>\n",
       "      <td>2.7</td>\n",
       "      <td>2.6</td>\n",
       "      <td>4.9</td>\n",
       "      <td>5.8</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-02-01</td>\n",
       "      <td>4.7</td>\n",
       "      <td>6.3</td>\n",
       "      <td>4.1</td>\n",
       "      <td>4.3</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.8</td>\n",
       "      <td>2.7</td>\n",
       "      <td>3.6</td>\n",
       "      <td>3.7</td>\n",
       "      <td>3.6</td>\n",
       "      <td>...</td>\n",
       "      <td>2.4</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.6</td>\n",
       "      <td>3.1</td>\n",
       "      <td>2.6</td>\n",
       "      <td>2.5</td>\n",
       "      <td>4.9</td>\n",
       "      <td>5.6</td>\n",
       "      <td>3.2</td>\n",
       "      <td>3.9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>4.6</td>\n",
       "      <td>6.3</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.3</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.7</td>\n",
       "      <td>2.6</td>\n",
       "      <td>3.6</td>\n",
       "      <td>3.7</td>\n",
       "      <td>3.6</td>\n",
       "      <td>...</td>\n",
       "      <td>2.4</td>\n",
       "      <td>3.8</td>\n",
       "      <td>4.5</td>\n",
       "      <td>3.1</td>\n",
       "      <td>2.6</td>\n",
       "      <td>2.4</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.5</td>\n",
       "      <td>3.3</td>\n",
       "      <td>3.9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-04-01</td>\n",
       "      <td>4.6</td>\n",
       "      <td>6.3</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.3</td>\n",
       "      <td>5.1</td>\n",
       "      <td>2.7</td>\n",
       "      <td>2.5</td>\n",
       "      <td>3.7</td>\n",
       "      <td>3.7</td>\n",
       "      <td>3.7</td>\n",
       "      <td>...</td>\n",
       "      <td>2.4</td>\n",
       "      <td>3.8</td>\n",
       "      <td>4.4</td>\n",
       "      <td>3.1</td>\n",
       "      <td>2.7</td>\n",
       "      <td>2.4</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.4</td>\n",
       "      <td>3.4</td>\n",
       "      <td>3.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-05-01</td>\n",
       "      <td>4.5</td>\n",
       "      <td>6.3</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.2</td>\n",
       "      <td>5.1</td>\n",
       "      <td>2.7</td>\n",
       "      <td>2.4</td>\n",
       "      <td>3.7</td>\n",
       "      <td>3.7</td>\n",
       "      <td>3.7</td>\n",
       "      <td>...</td>\n",
       "      <td>2.4</td>\n",
       "      <td>3.9</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.2</td>\n",
       "      <td>2.7</td>\n",
       "      <td>2.3</td>\n",
       "      <td>5.1</td>\n",
       "      <td>5.4</td>\n",
       "      <td>3.5</td>\n",
       "      <td>3.8</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 50 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "state       Alabama  Alaska  Arizona  Arkansas  California  Colorado  \\\n",
       "Date                                                                   \n",
       "2000-01-01      4.7     6.3      4.1       4.4         5.0       2.8   \n",
       "2000-02-01      4.7     6.3      4.1       4.3         5.0       2.8   \n",
       "2000-03-01      4.6     6.3      4.0       4.3         5.0       2.7   \n",
       "2000-04-01      4.6     6.3      4.0       4.3         5.1       2.7   \n",
       "2000-05-01      4.5     6.3      4.0       4.2         5.1       2.7   \n",
       "\n",
       "state       Connecticut  Delaware  Florida  Georgia  ...  South Dakota  \\\n",
       "Date                                                 ...                 \n",
       "2000-01-01          2.8       3.5      3.7      3.7  ...           2.4   \n",
       "2000-02-01          2.7       3.6      3.7      3.6  ...           2.4   \n",
       "2000-03-01          2.6       3.6      3.7      3.6  ...           2.4   \n",
       "2000-04-01          2.5       3.7      3.7      3.7  ...           2.4   \n",
       "2000-05-01          2.4       3.7      3.7      3.7  ...           2.4   \n",
       "\n",
       "state       Tennessee  Texas  Utah  Vermont  Virginia  Washington  \\\n",
       "Date                                                                \n",
       "2000-01-01        3.7    4.6   3.1      2.7       2.6         4.9   \n",
       "2000-02-01        3.7    4.6   3.1      2.6       2.5         4.9   \n",
       "2000-03-01        3.8    4.5   3.1      2.6       2.4         5.0   \n",
       "2000-04-01        3.8    4.4   3.1      2.7       2.4         5.0   \n",
       "2000-05-01        3.9    4.3   3.2      2.7       2.3         5.1   \n",
       "\n",
       "state       West Virginia  Wisconsin  Wyoming  \n",
       "Date                                           \n",
       "2000-01-01            5.8        3.2      4.1  \n",
       "2000-02-01            5.6        3.2      3.9  \n",
       "2000-03-01            5.5        3.3      3.9  \n",
       "2000-04-01            5.4        3.4      3.8  \n",
       "2000-05-01            5.4        3.5      3.8  \n",
       "\n",
       "[5 rows x 50 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Don't worry about the details here quite yet\n",
    "unemp_all = (\n",
    "    unemp_raw\n",
    "    .reset_index()\n",
    "    .pivot_table(index=\"Date\", columns=\"state\", values=\"UnemploymentRate\")\n",
    ")\n",
    "unemp_all.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "Finally, we can filter it to look at a subset of the columns (i.e. “state” in this case)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-02-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-04-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-05-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.5</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state       Arizona  California  Florida  Illinois  Michigan  New York  Texas\n",
       "Date                                                                         \n",
       "2000-01-01      4.1         5.0      3.7       4.2       3.3       4.7    4.6\n",
       "2000-02-01      4.1         5.0      3.7       4.2       3.2       4.7    4.6\n",
       "2000-03-01      4.0         5.0      3.7       4.3       3.2       4.6    4.5\n",
       "2000-04-01      4.0         5.1      3.7       4.3       3.3       4.6    4.4\n",
       "2000-05-01      4.0         5.1      3.7       4.3       3.5       4.6    4.3"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "states = [\n",
    "    \"Arizona\", \"California\", \"Florida\", \"Illinois\",\n",
    "    \"Michigan\", \"New York\", \"Texas\"\n",
    "]\n",
    "unemp = unemp_all[states]\n",
    "unemp.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "When plotting, a DataFrame knows the column and index names."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fda108f3850>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "unemp.plot(figsize=(8, 6))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Dates in pandas\n",
    "\n",
    "You might have noticed that our index now has a nice format for the\n",
    "dates (`YYYY-MM-DD`) rather than just a year.\n",
    "\n",
    "This is because the `dtype` of the index is a variant of `datetime`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DatetimeIndex(['2000-01-01', '2000-02-01', '2000-03-01', '2000-04-01',\n",
       "               '2000-05-01', '2000-06-01', '2000-07-01', '2000-08-01',\n",
       "               '2000-09-01', '2000-10-01',\n",
       "               ...\n",
       "               '2017-03-01', '2017-04-01', '2017-05-01', '2017-06-01',\n",
       "               '2017-07-01', '2017-08-01', '2017-09-01', '2017-10-01',\n",
       "               '2017-11-01', '2017-12-01'],\n",
       "              dtype='datetime64[ns]', name='Date', length=216, freq=None)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp.index"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "We can index into a DataFrame with a `DatetimeIndex` using string\n",
    "representations of dates.\n",
    "\n",
    "For example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "state\n",
       "Arizona       4.1\n",
       "California    5.0\n",
       "Florida       3.7\n",
       "Illinois      4.2\n",
       "Michigan      3.3\n",
       "New York      4.7\n",
       "Texas         4.6\n",
       "Name: 2000-01-01 00:00:00, dtype: float64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Data corresponding to a single date\n",
    "unemp.loc[\"2000-01-01\", :]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-02-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-04-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-05-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.5</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-06-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.8</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state       Arizona  California  Florida  Illinois  Michigan  New York  Texas\n",
       "Date                                                                         \n",
       "2000-01-01      4.1         5.0      3.7       4.2       3.3       4.7    4.6\n",
       "2000-02-01      4.1         5.0      3.7       4.2       3.2       4.7    4.6\n",
       "2000-03-01      4.0         5.0      3.7       4.3       3.2       4.6    4.5\n",
       "2000-04-01      4.0         5.1      3.7       4.3       3.3       4.6    4.4\n",
       "2000-05-01      4.0         5.1      3.7       4.3       3.5       4.6    4.3\n",
       "2000-06-01      4.0         5.1      3.8       4.3       3.7       4.6    4.3"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Data for all days between New Years Day and June first in the year 2000\n",
    "unemp.loc[\"01/01/2000\":\"06/01/2000\", :]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "We will learn more about what pandas can do with dates and times in an\n",
    "upcoming lecture on time series data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## DataFrame Aggregations\n",
    "\n",
    "Let’s talk about *aggregations*.\n",
    "\n",
    "Loosely speaking, an aggregation is an operation that combines multiple\n",
    "values into a single value.\n",
    "\n",
    "For example, computing the mean of three numbers (for example\n",
    "`[0, 1, 2]`) returns a single number (1).\n",
    "\n",
    "We will use aggregations extensively as we analyze and manipulate our data.\n",
    "\n",
    "Thankfully, pandas makes this easy!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "### Built-in Aggregations\n",
    "\n",
    "pandas already has some of the most frequently used aggregations.\n",
    "\n",
    "For example:\n",
    "\n",
    "- Mean  (`mean`)  \n",
    "- Variance (`var`)  \n",
    "- Standard deviation (`std`)  \n",
    "- Minimum (`min`)  \n",
    "- Median (`median`)  \n",
    "- Maximum (`max`)  \n",
    "- etc…  \n",
    "\n",
    "\n",
    ">**Note**\n",
    ">\n",
    ">When looking for common operations, using “tab completion” goes a long way."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "state\n",
       "Arizona       6.301389\n",
       "California    7.299074\n",
       "Florida       6.048611\n",
       "Illinois      6.822685\n",
       "Michigan      7.492593\n",
       "New York      6.102315\n",
       "Texas         5.695370\n",
       "dtype: float64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp.mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "As seen above, the aggregation’s default is to aggregate each column.\n",
    "\n",
    "However, by using the `axis` keyword argument, you can do aggregations by\n",
    "row as well."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2000-01-01    0.352381\n",
       "2000-02-01    0.384762\n",
       "2000-03-01    0.364762\n",
       "2000-04-01    0.353333\n",
       "2000-05-01    0.294762\n",
       "dtype: float64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp.var(axis=1).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "### Writing Your Own Aggregation\n",
    "\n",
    "The built-in aggregations will get us pretty far in our analysis, but\n",
    "sometimes we need more flexibility.\n",
    "\n",
    "We can have pandas perform custom aggregations by following these two\n",
    "steps:\n",
    "\n",
    "1. Write a Python function that takes a `Series` as an input and\n",
    "  outputs a single value.  \n",
    "1. Call the `agg` method with our new function as an argument.  \n",
    "\n",
    "\n",
    "For example, below, we will classify states as “low unemployment” or\n",
    "“high unemployment” based on whether their mean unemployment level is\n",
    "above or below 6.5."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "#\n",
    "# Step 1: We write the (aggregation) function that we'd like to use\n",
    "#\n",
    "def high_or_low(s):\n",
    "    \"\"\"\n",
    "    This function takes a pandas Series object and returns high\n",
    "    if the mean is above 6.5 and low if the mean is below 6.5\n",
    "    \"\"\"\n",
    "    if s.mean() < 6.5:\n",
    "        out = \"Low\"\n",
    "    else:\n",
    "        out = \"High\"\n",
    "\n",
    "    return out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "state\n",
       "Arizona        Low\n",
       "California    High\n",
       "Florida        Low\n",
       "Illinois      High\n",
       "Michigan      High\n",
       "New York       Low\n",
       "Texas          Low\n",
       "dtype: object"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#\n",
    "# Step 2: Apply it via the agg method.\n",
    "#\n",
    "unemp.agg(high_or_low)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2000-01-01    Low\n",
       "2000-02-01    Low\n",
       "2000-03-01    Low\n",
       "2000-04-01    Low\n",
       "2000-05-01    Low\n",
       "dtype: object"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# How does this differ from unemp.agg(high_or_low)?\n",
    "unemp.agg(high_or_low, axis=1).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "Notice that `agg` can also accept multiple functions at once."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>min</td>\n",
       "      <td>3.6</td>\n",
       "      <td>4.5</td>\n",
       "      <td>3.1</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>max</td>\n",
       "      <td>10.9</td>\n",
       "      <td>12.3</td>\n",
       "      <td>11.3</td>\n",
       "      <td>11.3</td>\n",
       "      <td>14.6</td>\n",
       "      <td>8.9</td>\n",
       "      <td>8.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>high_or_low</td>\n",
       "      <td>Low</td>\n",
       "      <td>High</td>\n",
       "      <td>Low</td>\n",
       "      <td>High</td>\n",
       "      <td>High</td>\n",
       "      <td>Low</td>\n",
       "      <td>Low</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Arizona California Florida Illinois Michigan New York Texas\n",
       "min             3.6        4.5     3.1      4.2      3.2      4.2   3.9\n",
       "max            10.9       12.3    11.3     11.3     14.6      8.9   8.3\n",
       "high_or_low     Low       High     Low     High     High      Low   Low"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp.agg([min, max, high_or_low])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "**Exercise 2**\n",
    "\n",
    "Do the following exercises in separate code cells below:\n",
    "\n",
    "- At each date, what is the minimum unemployment rate across all states\n",
    "  in our sample?  \n",
    "- What was the median unemployment rate in each state?  \n",
    "- What was the maximum unemployment rate across the states in our\n",
    "  sample? What state did it happen in? In what month/year was this\n",
    "  achieved?  \n",
    "  - Hint 1: What Python type (not `dtype`) is returned by the\n",
    "    aggregation?  \n",
    "  - Hint 2: Read documentation for the method `idxmax`  \n",
    "- Classify each state as high or low volatility based on whether the\n",
    "  variance of their unemployment is above or below 4.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2000-01-01    3.3\n",
       "2000-02-01    3.2\n",
       "2000-03-01    3.2\n",
       "2000-04-01    3.3\n",
       "2000-05-01    3.5\n",
       "             ... \n",
       "2017-08-01    4.0\n",
       "2017-09-01    3.9\n",
       "2017-10-01    3.9\n",
       "2017-11-01    3.9\n",
       "2017-12-01    3.9\n",
       "Length: 216, dtype: float64"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp.min(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "state\n",
       "Arizona       5.80\n",
       "California    6.50\n",
       "Florida       5.35\n",
       "Illinois      6.15\n",
       "Michigan      7.00\n",
       "New York      5.70\n",
       "Texas         5.40\n",
       "dtype: float64"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp.median()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "14.6\n",
      "Michigan\n",
      "2009-06-01 00:00:00\n"
     ]
    }
   ],
   "source": [
    "print(unemp.max().max())\n",
    "print(unemp.max().idxmax())\n",
    "print(unemp.max(axis=1).idxmax())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "state\n",
       "Arizona       high\n",
       "California    high\n",
       "Florida       high\n",
       "Illinois       low\n",
       "Michigan      high\n",
       "New York       low\n",
       "Texas          low\n",
       "dtype: object"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def ur_volatility_classify(s):\n",
    "    \"\"\"\n",
    "    Classifies the volatility of a series. A series is deemed\n",
    "    high volatility if the variance is higher than 4\n",
    "    \n",
    "    Parameters\n",
    "    ----------\n",
    "    s : Series\n",
    "        The series that we are classifying\n",
    "    \n",
    "    Returns\n",
    "    -------\n",
    "    _ : str\n",
    "        Either \"high\" or \"low\"\n",
    "    \"\"\"\n",
    "    var = s.var()\n",
    "    \n",
    "    return \"high\" if var > 4 else \"low\"\n",
    "\n",
    "unemp.apply(ur_volatility_classify)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Transforms\n",
    "\n",
    "Many analytical operations do not necessarily involve an aggregation.\n",
    "\n",
    "The output of a function applied to a Series might need to be a new\n",
    "Series.\n",
    "\n",
    "Some examples:\n",
    "\n",
    "- Compute the percentage change in unemployment from month to month.  \n",
    "- Calculate the cumulative sum of elements in each column.  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Built-in Transforms\n",
    "\n",
    "pandas comes with many transform functions including:\n",
    "\n",
    "- Cumulative sum/max/min/product (`cum(sum|min|max|prod)`)  \n",
    "- Difference  (`diff`)  \n",
    "- Elementwise addition/subtraction/multiplication/division (`+`, `-`, `*`, `/`)  \n",
    "- Percent change (`pct_change`)  \n",
    "- Number of occurrences of each distinct value (`value_counts`)  \n",
    "- Absolute value (`abs`)  \n",
    "\n",
    "\n",
    "Again, tab completion is helpful when trying to find these functions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-02-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-04-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-05-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.5</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state       Arizona  California  Florida  Illinois  Michigan  New York  Texas\n",
       "Date                                                                         \n",
       "2000-01-01      4.1         5.0      3.7       4.2       3.3       4.7    4.6\n",
       "2000-02-01      4.1         5.0      3.7       4.2       3.2       4.7    4.6\n",
       "2000-03-01      4.0         5.0      3.7       4.3       3.2       4.6    4.5\n",
       "2000-04-01      4.0         5.1      3.7       4.3       3.3       4.6    4.4\n",
       "2000-05-01      4.0         5.1      3.7       4.3       3.5       4.6    4.3"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-02-01</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>-0.030303</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>-0.02439</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.02381</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.021277</td>\n",
       "      <td>-0.021739</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-04-01</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.02</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.031250</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.022222</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-05-01</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.060606</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.022727</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state       Arizona  California  Florida  Illinois  Michigan  New York  \\\n",
       "Date                                                                     \n",
       "2000-01-01      NaN         NaN      NaN       NaN       NaN       NaN   \n",
       "2000-02-01  0.00000        0.00      0.0   0.00000 -0.030303  0.000000   \n",
       "2000-03-01 -0.02439        0.00      0.0   0.02381  0.000000 -0.021277   \n",
       "2000-04-01  0.00000        0.02      0.0   0.00000  0.031250  0.000000   \n",
       "2000-05-01  0.00000        0.00      0.0   0.00000  0.060606  0.000000   \n",
       "\n",
       "state          Texas  \n",
       "Date                  \n",
       "2000-01-01       NaN  \n",
       "2000-02-01  0.000000  \n",
       "2000-03-01 -0.021739  \n",
       "2000-04-01 -0.022222  \n",
       "2000-05-01 -0.022727  "
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp.pct_change().head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-02-01</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>-0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.1</td>\n",
       "      <td>-0.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-04-01</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-05-01</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state       Arizona  California  Florida  Illinois  Michigan  New York  Texas\n",
       "Date                                                                         \n",
       "2000-01-01      NaN         NaN      NaN       NaN       NaN       NaN    NaN\n",
       "2000-02-01      0.0         0.0      0.0       0.0      -0.1       0.0    0.0\n",
       "2000-03-01     -0.1         0.0      0.0       0.1       0.0      -0.1   -0.1\n",
       "2000-04-01      0.0         0.1      0.0       0.0       0.1       0.0   -0.1\n",
       "2000-05-01      0.0         0.0      0.0       0.0       0.2       0.0   -0.1"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp.diff().head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "Transforms can be split into to several main categories:\n",
    "\n",
    "1. *Series transforms*: functions that take in one Series and produce another Series. The index of the input and output does not need to be the same.  \n",
    "1. *Scalar transforms*: functions that take a single value and produce a single value. An example is the `abs` method, or adding a constant to each value of a Series.  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "### Custom Series Transforms\n",
    "\n",
    "pandas also simplifies applying custom Series transforms to a Series or the\n",
    "columns of a DataFrame. The steps are:\n",
    "\n",
    "1. Write a Python function that takes a Series and outputs a new Series.  \n",
    "1. Pass our new function as an argument to the `apply` method (alternatively, the `transform` method).  \n",
    "\n",
    "\n",
    "As an example, we will standardize our unemployment data to have mean 0\n",
    "and standard deviation 1.\n",
    "\n",
    "After doing this, we can use an aggregation to determine at which date the\n",
    "unemployment rate is most different from “normal times” for each state."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "#\n",
    "# Step 1: We write the Series transform function that we'd like to use\n",
    "#\n",
    "def standardize_data(x):\n",
    "    \"\"\"\n",
    "    Changes the data in a Series to become mean 0 with standard deviation 1\n",
    "    \"\"\"\n",
    "    mu = x.mean()\n",
    "    std = x.std()\n",
    "\n",
    "    return (x - mu)/std"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>-1.076861</td>\n",
       "      <td>-0.935545</td>\n",
       "      <td>-0.976846</td>\n",
       "      <td>-1.337203</td>\n",
       "      <td>-1.605740</td>\n",
       "      <td>-0.925962</td>\n",
       "      <td>-0.849345</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-02-01</td>\n",
       "      <td>-1.076861</td>\n",
       "      <td>-0.935545</td>\n",
       "      <td>-0.976846</td>\n",
       "      <td>-1.337203</td>\n",
       "      <td>-1.644039</td>\n",
       "      <td>-0.925962</td>\n",
       "      <td>-0.849345</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>-1.125778</td>\n",
       "      <td>-0.935545</td>\n",
       "      <td>-0.976846</td>\n",
       "      <td>-1.286217</td>\n",
       "      <td>-1.644039</td>\n",
       "      <td>-0.991993</td>\n",
       "      <td>-0.926885</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-04-01</td>\n",
       "      <td>-1.125778</td>\n",
       "      <td>-0.894853</td>\n",
       "      <td>-0.976846</td>\n",
       "      <td>-1.286217</td>\n",
       "      <td>-1.605740</td>\n",
       "      <td>-0.991993</td>\n",
       "      <td>-1.004424</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-05-01</td>\n",
       "      <td>-1.125778</td>\n",
       "      <td>-0.894853</td>\n",
       "      <td>-0.976846</td>\n",
       "      <td>-1.286217</td>\n",
       "      <td>-1.529141</td>\n",
       "      <td>-0.991993</td>\n",
       "      <td>-1.081964</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state        Arizona  California   Florida  Illinois  Michigan  New York  \\\n",
       "Date                                                                       \n",
       "2000-01-01 -1.076861   -0.935545 -0.976846 -1.337203 -1.605740 -0.925962   \n",
       "2000-02-01 -1.076861   -0.935545 -0.976846 -1.337203 -1.644039 -0.925962   \n",
       "2000-03-01 -1.125778   -0.935545 -0.976846 -1.286217 -1.644039 -0.991993   \n",
       "2000-04-01 -1.125778   -0.894853 -0.976846 -1.286217 -1.605740 -0.991993   \n",
       "2000-05-01 -1.125778   -0.894853 -0.976846 -1.286217 -1.529141 -0.991993   \n",
       "\n",
       "state          Texas  \n",
       "Date                  \n",
       "2000-01-01 -0.849345  \n",
       "2000-02-01 -0.849345  \n",
       "2000-03-01 -0.926885  \n",
       "2000-04-01 -1.004424  \n",
       "2000-05-01 -1.081964  "
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#\n",
    "# Step 2: Apply our function via the apply method.\n",
    "#\n",
    "std_unemp = unemp.apply(standardize_data)\n",
    "std_unemp.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>1.076861</td>\n",
       "      <td>0.935545</td>\n",
       "      <td>0.976846</td>\n",
       "      <td>1.337203</td>\n",
       "      <td>1.605740</td>\n",
       "      <td>0.925962</td>\n",
       "      <td>0.849345</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-02-01</td>\n",
       "      <td>1.076861</td>\n",
       "      <td>0.935545</td>\n",
       "      <td>0.976846</td>\n",
       "      <td>1.337203</td>\n",
       "      <td>1.644039</td>\n",
       "      <td>0.925962</td>\n",
       "      <td>0.849345</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>1.125778</td>\n",
       "      <td>0.935545</td>\n",
       "      <td>0.976846</td>\n",
       "      <td>1.286217</td>\n",
       "      <td>1.644039</td>\n",
       "      <td>0.991993</td>\n",
       "      <td>0.926885</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-04-01</td>\n",
       "      <td>1.125778</td>\n",
       "      <td>0.894853</td>\n",
       "      <td>0.976846</td>\n",
       "      <td>1.286217</td>\n",
       "      <td>1.605740</td>\n",
       "      <td>0.991993</td>\n",
       "      <td>1.004424</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-05-01</td>\n",
       "      <td>1.125778</td>\n",
       "      <td>0.894853</td>\n",
       "      <td>0.976846</td>\n",
       "      <td>1.286217</td>\n",
       "      <td>1.529141</td>\n",
       "      <td>0.991993</td>\n",
       "      <td>1.081964</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state        Arizona  California   Florida  Illinois  Michigan  New York  \\\n",
       "Date                                                                       \n",
       "2000-01-01  1.076861    0.935545  0.976846  1.337203  1.605740  0.925962   \n",
       "2000-02-01  1.076861    0.935545  0.976846  1.337203  1.644039  0.925962   \n",
       "2000-03-01  1.125778    0.935545  0.976846  1.286217  1.644039  0.991993   \n",
       "2000-04-01  1.125778    0.894853  0.976846  1.286217  1.605740  0.991993   \n",
       "2000-05-01  1.125778    0.894853  0.976846  1.286217  1.529141  0.991993   \n",
       "\n",
       "state          Texas  \n",
       "Date                  \n",
       "2000-01-01  0.849345  \n",
       "2000-02-01  0.849345  \n",
       "2000-03-01  0.926885  \n",
       "2000-04-01  1.004424  \n",
       "2000-05-01  1.081964  "
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Takes the absolute value of all elements of a function\n",
    "abs_std_unemp = std_unemp.abs()\n",
    "\n",
    "abs_std_unemp.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "state\n",
       "Arizona      2009-11-01\n",
       "California   2010-03-01\n",
       "Florida      2010-01-01\n",
       "Illinois     2009-12-01\n",
       "Michigan     2009-06-01\n",
       "New York     2009-11-01\n",
       "Texas        2009-08-01\n",
       "dtype: datetime64[ns]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# find the date when unemployment was \"most different from normal\" for each State\n",
    "def idxmax(x):\n",
    "    # idxmax of Series will return index of maximal value\n",
    "    return x.idxmax()\n",
    "\n",
    "abs_std_unemp.agg(idxmax)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Custom Scalar Transforms\n",
    "\n",
    "As you may have predicted, we can also apply custom scalar transforms to our\n",
    "pandas data.\n",
    "\n",
    "To do this, we use the following pattern:\n",
    "\n",
    "1. Define a Python function that takes in a scalar and produces a scalar.  \n",
    "1. Pass this function as an argument to the `applymap` Series or DataFrame method.  \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "**Example**\n",
    "\n",
    "Imagine that we want to determine whether unemployment was high (> 6.5),\n",
    "medium (4.5 < x <= 6.5), or low (<= 4.5) for each state and each month.\n",
    "\n",
    "1. Write a Python function that takes a single number as an input and\n",
    "  outputs a single string noting if that number is high, medium, or low.  \n",
    "2. Pass your function to `applymap` (quiz: why `applymap` and not\n",
    "  `agg` or `apply`?) and save the result in a new DataFrame called\n",
    "  `unemp_bins`.  \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "def unemployment_classifier(ur):\n",
    "    \"\"\"\n",
    "    Classifies the unemployment rate as high, medium, or low\n",
    "    based on the value\n",
    "    \n",
    "    Parameters\n",
    "    ----------\n",
    "    ur : scalar(float)\n",
    "        The unemployment rate\n",
    "    \n",
    "    Returns\n",
    "    -------\n",
    "    out : str\n",
    "        The classification \"high\", \"medium\", or \"low\"\n",
    "    \"\"\"\n",
    "    if ur > 6.5:\n",
    "        return \"high\"\n",
    "    elif ur > 4.5:\n",
    "        return \"medium\"\n",
    "    else:\n",
    "        return \"low\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>low</td>\n",
       "      <td>medium</td>\n",
       "      <td>low</td>\n",
       "      <td>low</td>\n",
       "      <td>low</td>\n",
       "      <td>medium</td>\n",
       "      <td>medium</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-02-01</td>\n",
       "      <td>low</td>\n",
       "      <td>medium</td>\n",
       "      <td>low</td>\n",
       "      <td>low</td>\n",
       "      <td>low</td>\n",
       "      <td>medium</td>\n",
       "      <td>medium</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>low</td>\n",
       "      <td>medium</td>\n",
       "      <td>low</td>\n",
       "      <td>low</td>\n",
       "      <td>low</td>\n",
       "      <td>medium</td>\n",
       "      <td>low</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-04-01</td>\n",
       "      <td>low</td>\n",
       "      <td>medium</td>\n",
       "      <td>low</td>\n",
       "      <td>low</td>\n",
       "      <td>low</td>\n",
       "      <td>medium</td>\n",
       "      <td>low</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-05-01</td>\n",
       "      <td>low</td>\n",
       "      <td>medium</td>\n",
       "      <td>low</td>\n",
       "      <td>low</td>\n",
       "      <td>low</td>\n",
       "      <td>medium</td>\n",
       "      <td>low</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state      Arizona California Florida Illinois Michigan New York   Texas\n",
       "Date                                                                    \n",
       "2000-01-01     low     medium     low      low      low   medium  medium\n",
       "2000-02-01     low     medium     low      low      low   medium  medium\n",
       "2000-03-01     low     medium     low      low      low   medium     low\n",
       "2000-04-01     low     medium     low      low      low   medium     low\n",
       "2000-05-01     low     medium     low      low      low   medium     low"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp_bins = unemp.applymap(unemployment_classifier)\n",
    "unemp_bins.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "3. (Challenging) This exercise has multiple parts:  \n",
    "  1. Use another transform on `unemp_bins` to count how many\n",
    "    times each state had each of the three classifications.  \n",
    "    - Hint 1: Will this value counting function be a Series or scalar\n",
    "      transform?  \n",
    "    - Hint 2: Try googling \"pandas count unique value\" or something\n",
    "      similar to find the right transform.  \n",
    "  2. Construct a horizontal bar chart of the number of occurrences of\n",
    "    each level with one bar per state and classification (21 total\n",
    "    bars).  \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>high</td>\n",
       "      <td>75</td>\n",
       "      <td>106</td>\n",
       "      <td>68</td>\n",
       "      <td>91</td>\n",
       "      <td>142</td>\n",
       "      <td>65</td>\n",
       "      <td>51</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>low</td>\n",
       "      <td>44</td>\n",
       "      <td>4</td>\n",
       "      <td>69</td>\n",
       "      <td>19</td>\n",
       "      <td>17</td>\n",
       "      <td>22</td>\n",
       "      <td>58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>medium</td>\n",
       "      <td>97</td>\n",
       "      <td>106</td>\n",
       "      <td>79</td>\n",
       "      <td>106</td>\n",
       "      <td>57</td>\n",
       "      <td>129</td>\n",
       "      <td>107</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state   Arizona  California  Florida  Illinois  Michigan  New York  Texas\n",
       "high         75         106       68        91       142        65     51\n",
       "low          44           4       69        19        17        22     58\n",
       "medium       97         106       79       106        57       129    107"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp_class_count = unemp_bins.apply(\n",
    "    pd.Series.value_counts\n",
    ")\n",
    "\n",
    "unemp_class_count.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fda0c2a8510>"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "unemp_class_count.T.loc[:, [\"low\", \"medium\", \"high\"]].plot(kind=\"bar\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "4. (Challenging) Repeat the previous step, but count how many states had\n",
    "  each classification in each month. Which month had the most states\n",
    "  with high unemployment? What about medium and low?  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>high</th>\n",
       "      <th>low</th>\n",
       "      <th>medium</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-02-01</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-04-01</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-05-01</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            high  low  medium\n",
       "Date                         \n",
       "2000-01-01   0.0  4.0     3.0\n",
       "2000-02-01   0.0  4.0     3.0\n",
       "2000-03-01   0.0  5.0     2.0\n",
       "2000-04-01   0.0  5.0     2.0\n",
       "2000-05-01   0.0  5.0     2.0"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp_monthly_count = unemp_bins.apply(\n",
    "    pd.Series.value_counts, axis=1\n",
    ").fillna(0.0)\n",
    "\n",
    "unemp_monthly_count.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Timestamp('2009-04-01 00:00:00')"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp_monthly_count[\"high\"].idxmax()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Timestamp('2001-09-01 00:00:00')"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp_monthly_count[\"medium\"].idxmax()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Timestamp('2000-08-01 00:00:00')"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp_monthly_count[\"low\"].idxmax()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Boolean Selection\n",
    "\n",
    "We have seen how we can select subsets of data by referring to the index\n",
    "or column names.\n",
    "\n",
    "However, we often want to select based on conditions met by\n",
    "the data itself.\n",
    "\n",
    "Some examples are:\n",
    "\n",
    "- Restrict analysis to all individuals older than 18.  \n",
    "- Look at data that corresponds to particular time periods.  \n",
    "- Analyze only data that corresponds to a recession.  \n",
    "- Obtain data for a specific product or customer ID.  \n",
    "\n",
    "\n",
    "We will be able to do this by using a Series or list of boolean values\n",
    "to index into a Series or DataFrame.\n",
    "\n",
    "Let’s look at some examples."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-02-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-04-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-05-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.5</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state       Arizona  California  Florida  Illinois  Michigan  New York  Texas\n",
       "Date                                                                         \n",
       "2000-01-01      4.1         5.0      3.7       4.2       3.3       4.7    4.6\n",
       "2000-02-01      4.1         5.0      3.7       4.2       3.2       4.7    4.6\n",
       "2000-03-01      4.0         5.0      3.7       4.3       3.2       4.6    4.5\n",
       "2000-04-01      4.0         5.1      3.7       4.3       3.3       4.6    4.4\n",
       "2000-05-01      4.0         5.1      3.7       4.3       3.5       4.6    4.3"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp_small = unemp.head()  # Create smaller data so we can see what's happening\n",
    "unemp_small"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-02-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state       Arizona  California  Florida  Illinois  Michigan  New York  Texas\n",
       "Date                                                                         \n",
       "2000-01-01      4.1         5.0      3.7       4.2       3.3       4.7    4.6\n",
       "2000-02-01      4.1         5.0      3.7       4.2       3.2       4.7    4.6\n",
       "2000-03-01      4.0         5.0      3.7       4.3       3.2       4.6    4.5"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# list of booleans selects rows\n",
    "unemp_small.loc[[True, True, True, False, False], :]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-05-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.5</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state       Arizona  California  Florida  Illinois  Michigan  New York  Texas\n",
       "Date                                                                         \n",
       "2000-01-01      4.1         5.0      3.7       4.2       3.3       4.7    4.6\n",
       "2000-03-01      4.0         5.0      3.7       4.3       3.2       4.6    4.5\n",
       "2000-05-01      4.0         5.1      3.7       4.3       3.5       4.6    4.3"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# second argument selects columns, the  ``:``  means \"all\".\n",
    "# here we use it to select all columns\n",
    "unemp_small.loc[[True, False, True, False, True], :]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-02-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state       Arizona  New York  Texas\n",
       "Date                                \n",
       "2000-01-01      4.1       4.7    4.6\n",
       "2000-02-01      4.1       4.7    4.6\n",
       "2000-03-01      4.0       4.6    4.5"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# can use booleans to select both rows and columns\n",
    "unemp_small.loc[[True, True, True, False, False], [True, False, False, False, False, True, True]]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "### Creating Boolean DataFrames/Series\n",
    "\n",
    "We can use [conditional statements](../python_fundamentals/control_flow.ipynb) to\n",
    "construct Series of booleans from our data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2000-01-01    False\n",
       "2000-02-01    False\n",
       "2000-03-01    False\n",
       "2000-04-01     True\n",
       "2000-05-01     True\n",
       "Name: Texas, dtype: bool"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp_small[\"Texas\"] < 4.5"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "Once we have our Series of bools, we can use it to extract subsets of\n",
    "rows from our DataFrame."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-04-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-05-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.5</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state       Arizona  California  Florida  Illinois  Michigan  New York  Texas\n",
       "Date                                                                         \n",
       "2000-04-01      4.0         5.1      3.7       4.3       3.3       4.6    4.4\n",
       "2000-05-01      4.0         5.1      3.7       4.3       3.5       4.6    4.3"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp_small.loc[unemp_small[\"Texas\"] < 4.5, :]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2000-01-01    True\n",
       "2000-02-01    True\n",
       "2000-03-01    True\n",
       "2000-04-01    True\n",
       "2000-05-01    True\n",
       "dtype: bool"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp_small[\"New York\"] > unemp_small[\"Texas\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-02-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-04-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-05-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.5</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state       Arizona  California  Florida  Illinois  Michigan  New York  Texas\n",
       "Date                                                                         \n",
       "2000-01-01      4.1         5.0      3.7       4.2       3.3       4.7    4.6\n",
       "2000-02-01      4.1         5.0      3.7       4.2       3.2       4.7    4.6\n",
       "2000-03-01      4.0         5.0      3.7       4.3       3.2       4.6    4.5\n",
       "2000-04-01      4.0         5.1      3.7       4.3       3.3       4.6    4.4\n",
       "2000-05-01      4.0         5.1      3.7       4.3       3.5       4.6    4.3"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "big_NY = unemp_small[\"New York\"] > unemp_small[\"Texas\"]\n",
    "unemp_small.loc[big_NY]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "#### Multiple Conditions\n",
    "\n",
    "In the boolean section of the [basics lecture](../python_fundamentals/basics.ipynb), we saw\n",
    "that we can use the words `and` and `or` to combine multiple booleans into\n",
    "a single bool.\n",
    "\n",
    "Recall:\n",
    "\n",
    "- `True and False -> False`  \n",
    "- `True and True -> True`  \n",
    "- `False and False -> False`  \n",
    "- `True or False -> True`  \n",
    "- `True or True -> True`  \n",
    "- `False or False -> False`  \n",
    "\n",
    "\n",
    "We can do something similar in pandas, but instead of\n",
    "`bool1 and bool2` we write:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "```python\n",
    "(bool_series1) & (bool_series2)\n",
    "```\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "Likewise, instead of `bool1 or bool2` we write:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "```python\n",
    "(bool_series1) | (bool_series2)\n",
    "```\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2000-01-01    False\n",
       "2000-02-01    False\n",
       "2000-03-01     True\n",
       "2000-04-01     True\n",
       "2000-05-01     True\n",
       "dtype: bool"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "small_NYTX = (unemp_small[\"Texas\"] < 4.7) & (unemp_small[\"New York\"] < 4.7)\n",
    "small_NYTX"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-04-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-05-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.5</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state       Arizona  California  Florida  Illinois  Michigan  New York  Texas\n",
       "Date                                                                         \n",
       "2000-03-01      4.0         5.0      3.7       4.3       3.2       4.6    4.5\n",
       "2000-04-01      4.0         5.1      3.7       4.3       3.3       4.6    4.4\n",
       "2000-05-01      4.0         5.1      3.7       4.3       3.5       4.6    4.3"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp_small[small_NYTX]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "#### `isin`\n",
    "\n",
    "Sometimes, we will want to check whether a data point takes on one of a\n",
    "several fixed values.\n",
    "\n",
    "We could do this by writing `(df[\"x\"] == val_1) | (df[\"x\"] == val_2)`\n",
    "(like we did above), but there is a better way: the `.isin` method"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2000-01-01     True\n",
       "2000-02-01     True\n",
       "2000-03-01     True\n",
       "2000-04-01     True\n",
       "2000-05-01    False\n",
       "Name: Michigan, dtype: bool"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unemp_small[\"Michigan\"].isin([3.3, 3.2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-02-01</td>\n",
       "      <td>4.1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.2</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.7</td>\n",
       "      <td>4.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.2</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-04-01</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>3.7</td>\n",
       "      <td>4.3</td>\n",
       "      <td>3.3</td>\n",
       "      <td>4.6</td>\n",
       "      <td>4.4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state       Arizona  California  Florida  Illinois  Michigan  New York  Texas\n",
       "Date                                                                         \n",
       "2000-01-01      4.1         5.0      3.7       4.2       3.3       4.7    4.6\n",
       "2000-02-01      4.1         5.0      3.7       4.2       3.2       4.7    4.6\n",
       "2000-03-01      4.0         5.0      3.7       4.3       3.2       4.6    4.5\n",
       "2000-04-01      4.0         5.1      3.7       4.3       3.3       4.6    4.4"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# now select full rows where this Series is True\n",
    "unemp_small.loc[unemp_small[\"Michigan\"].isin([3.3, 3.2])]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "#### `.any` and `.all`\n",
    "\n",
    "Recall from the boolean section of the [basics lecture](../python_fundamentals/basics.ipynb)\n",
    "that the Python functions `any` and `all` are aggregation functions that\n",
    "take a collection of booleans and return a single boolean.\n",
    "\n",
    "`any` returns True whenever at least one of the inputs are True while\n",
    "`all` is True only when all the inputs are `True`.\n",
    "\n",
    "Series and DataFrames with `dtype` bool have `.any` and `.all`\n",
    "methods that apply this logic to pandas objects.\n",
    "\n",
    "Let’s use these methods to count how many months all the states in our\n",
    "sample had high unemployment.\n",
    "\n",
    "As we work through this example, consider the [“want\n",
    "operator”](http://albertjmenkveld.com/2014/07/07/endogeneous-price-dispersion/), a helpful\n",
    "concept we learned from [Tom\n",
    "Sargent](http://www.tomsargent.com) for clearly stating the goal of our analysis and\n",
    "determining the steps necessary to reach the goal.\n",
    "\n",
    "We always begin by writing `Want:` followed by what we want to\n",
    "accomplish.\n",
    "\n",
    "In this case, we would write:\n",
    "\n",
    "> Want: Count the number of months in which all states in our sample\n",
    "had unemployment above 6.5%\n",
    "\n",
    "\n",
    "After identifying the **want**, we work *backwards* to identify the\n",
    "steps necessary to accomplish our goal.\n",
    "\n",
    "So, starting from the result, we have:\n",
    "\n",
    "1. Sum the number of `True` values in a Series indicating dates for\n",
    "  which all states had high unemployment.  \n",
    "1. Build the Series used in the last step by using the `.all` method\n",
    "  on a DataFrame containing booleans indicating whether each state had\n",
    "  high unemployment at each date.  \n",
    "1. Build the DataFrame used in the previous step using a `>`\n",
    "  comparison.  \n",
    "\n",
    "\n",
    "Now that we have a clear plan, let’s follow through and *apply* the want\n",
    "operator:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>state</th>\n",
       "      <th>Arizona</th>\n",
       "      <th>California</th>\n",
       "      <th>Florida</th>\n",
       "      <th>Illinois</th>\n",
       "      <th>Michigan</th>\n",
       "      <th>New York</th>\n",
       "      <th>Texas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2000-01-01</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-02-01</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-03-01</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-04-01</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2000-05-01</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "state       Arizona  California  Florida  Illinois  Michigan  New York  Texas\n",
       "Date                                                                         \n",
       "2000-01-01    False       False    False     False     False     False  False\n",
       "2000-02-01    False       False    False     False     False     False  False\n",
       "2000-03-01    False       False    False     False     False     False  False\n",
       "2000-04-01    False       False    False     False     False     False  False\n",
       "2000-05-01    False       False    False     False     False     False  False"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Step 3: construct the DataFrame of bools\n",
    "high = unemp > 6.5\n",
    "high.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date\n",
       "2000-01-01    False\n",
       "2000-02-01    False\n",
       "2000-03-01    False\n",
       "2000-04-01    False\n",
       "2000-05-01    False\n",
       "dtype: bool"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Step 2: use the .all method on axis=1 to get the dates where all states have a True\n",
    "all_high = high.all(axis=1)\n",
    "all_high.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "hide-output": false,
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Out of 216 months, 41 had high unemployment across all states\n"
     ]
    }
   ],
   "source": [
    "# Step 1: Call .sum to add up the number of True values in `all_high`\n",
    "#         (note that True == 1 and False == 0 in Python, so .sum will count Trues)\n",
    "msg = \"Out of {} months, {} had high unemployment across all states\"\n",
    "print(msg.format(len(all_high), all_high.sum()))"
   ]
  }
 ],
 "metadata": {
  "celltoolbar": "Slideshow",
  "date": 1595352471.6656835,
  "download_nb": false,
  "filename": "basics.rst",
  "filename_with_path": "pandas/basics",
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.3"
  },
  "title": "Basic Functionality"
 },
 "nbformat": 4,
 "nbformat_minor": 2
}