uber.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "ea5fc081-770c-47c6-9492-56613342d432",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.impute import SimpleImputer\n",
    "from sklearn.preprocessing import  OneHotEncoder, StandardScaler\n",
    "from sklearn.compose import ColumnTransformer\n",
    "from sklearn.cluster import KMeans, DBSCAN\n",
    "from sklearn.metrics import DistanceMetric\n",
    "import plotly.express as px\n",
    "import plotly.graph_objects as go\n",
    "import plotly.io as pio\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import warnings\n",
    "from functions import calculate_centroid\n",
    "\n",
    "warnings.filterwarnings(\"ignore\", category=DeprecationWarning) # to avoid deprecation warnings\n",
    "pio.renderers.default = \"vscode\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "c3db332b-ae12-400f-a163-27068cdbbb66",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('uber-raw-data-sep14.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "1be71343-788a-40b6-b06d-13b93e588fde",
   "metadata": {},
   "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/Time</th>\n",
       "      <th>Lat</th>\n",
       "      <th>Lon</th>\n",
       "      <th>Base</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>9/1/2014 0:01:00</td>\n",
       "      <td>40.2201</td>\n",
       "      <td>-74.0021</td>\n",
       "      <td>B02512</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>9/1/2014 0:01:00</td>\n",
       "      <td>40.7500</td>\n",
       "      <td>-74.0027</td>\n",
       "      <td>B02512</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>9/1/2014 0:03:00</td>\n",
       "      <td>40.7559</td>\n",
       "      <td>-73.9864</td>\n",
       "      <td>B02512</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>9/1/2014 0:06:00</td>\n",
       "      <td>40.7450</td>\n",
       "      <td>-73.9889</td>\n",
       "      <td>B02512</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>9/1/2014 0:11:00</td>\n",
       "      <td>40.8145</td>\n",
       "      <td>-73.9444</td>\n",
       "      <td>B02512</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Date/Time      Lat      Lon    Base\n",
       "0  9/1/2014 0:01:00  40.2201 -74.0021  B02512\n",
       "1  9/1/2014 0:01:00  40.7500 -74.0027  B02512\n",
       "2  9/1/2014 0:03:00  40.7559 -73.9864  B02512\n",
       "3  9/1/2014 0:06:00  40.7450 -73.9889  B02512\n",
       "4  9/1/2014 0:11:00  40.8145 -73.9444  B02512"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "03599417-2994-41ca-965c-0d199dbb2afb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1028136, 4)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "641119ad-3ade-4107-87ea-b9fff4e355b9",
   "metadata": {},
   "source": [
    "# Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "7c788add-3d70-4779-aff5-fa30f430ad9b",
   "metadata": {},
   "outputs": [],
   "source": [
    "df['Date/Time'] = pd.to_datetime(df['Date/Time'],format='%m/%d/%Y %H:%M:%S')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "4b001ef0-f79f-4101-a996-417319da5192",
   "metadata": {},
   "outputs": [],
   "source": [
    "df['Date/Time'] = pd.to_datetime(df['Date/Time'],format='%m/%d/%Y %H:%M:%S')\n",
    "df['Day'] = df['Date/Time'].dt.day\n",
    "df['DayOfWeek'] = df['Date/Time'].dt.dayofweek\n",
    "df['DayOfWeekLabel'] = df['Date/Time'].dt.day_name()\n",
    "df['Month'] = df['Date/Time'].dt.month\n",
    "df['Year'] = df['Date/Time'].dt.year\n",
    "df['Hour'] = df['Date/Time'].dt.hour\n",
    "df['Minute'] = df['Date/Time'].dt.minute"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "11199f8a",
   "metadata": {},
   "outputs": [],
   "source": [
    "df[\"DayOfWeekAndHour\"] = df[\"DayOfWeekLabel\"] + \"_\" + df[\"Hour\"].astype(str)+\"h\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "9cec7269-91db-464a-b103-c1b61bd9287d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Date/Time           0\n",
       "Lat                 0\n",
       "Lon                 0\n",
       "Base                0\n",
       "Day                 0\n",
       "DayOfWeek           0\n",
       "DayOfWeekLabel      0\n",
       "Month               0\n",
       "Year                0\n",
       "Hour                0\n",
       "Minute              0\n",
       "DayOfWeekAndHour    0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "2b338929",
   "metadata": {},
   "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/Time</th>\n",
       "      <th>Lat</th>\n",
       "      <th>Lon</th>\n",
       "      <th>Base</th>\n",
       "      <th>Day</th>\n",
       "      <th>DayOfWeek</th>\n",
       "      <th>DayOfWeekLabel</th>\n",
       "      <th>Month</th>\n",
       "      <th>Year</th>\n",
       "      <th>Hour</th>\n",
       "      <th>Minute</th>\n",
       "      <th>DayOfWeekAndHour</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2014-09-01 00:01:00</td>\n",
       "      <td>40.2201</td>\n",
       "      <td>-74.0021</td>\n",
       "      <td>B02512</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>Monday</td>\n",
       "      <td>9</td>\n",
       "      <td>2014</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>Monday_0h</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2014-09-01 00:01:00</td>\n",
       "      <td>40.7500</td>\n",
       "      <td>-74.0027</td>\n",
       "      <td>B02512</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>Monday</td>\n",
       "      <td>9</td>\n",
       "      <td>2014</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>Monday_0h</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2014-09-01 00:03:00</td>\n",
       "      <td>40.7559</td>\n",
       "      <td>-73.9864</td>\n",
       "      <td>B02512</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>Monday</td>\n",
       "      <td>9</td>\n",
       "      <td>2014</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>Monday_0h</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2014-09-01 00:06:00</td>\n",
       "      <td>40.7450</td>\n",
       "      <td>-73.9889</td>\n",
       "      <td>B02512</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>Monday</td>\n",
       "      <td>9</td>\n",
       "      <td>2014</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>Monday_0h</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2014-09-01 00:11:00</td>\n",
       "      <td>40.8145</td>\n",
       "      <td>-73.9444</td>\n",
       "      <td>B02512</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>Monday</td>\n",
       "      <td>9</td>\n",
       "      <td>2014</td>\n",
       "      <td>0</td>\n",
       "      <td>11</td>\n",
       "      <td>Monday_0h</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Date/Time      Lat      Lon    Base  Day  DayOfWeek  \\\n",
       "0 2014-09-01 00:01:00  40.2201 -74.0021  B02512    1          0   \n",
       "1 2014-09-01 00:01:00  40.7500 -74.0027  B02512    1          0   \n",
       "2 2014-09-01 00:03:00  40.7559 -73.9864  B02512    1          0   \n",
       "3 2014-09-01 00:06:00  40.7450 -73.9889  B02512    1          0   \n",
       "4 2014-09-01 00:11:00  40.8145 -73.9444  B02512    1          0   \n",
       "\n",
       "  DayOfWeekLabel  Month  Year  Hour  Minute DayOfWeekAndHour  \n",
       "0         Monday      9  2014     0       1        Monday_0h  \n",
       "1         Monday      9  2014     0       1        Monday_0h  \n",
       "2         Monday      9  2014     0       3        Monday_0h  \n",
       "3         Monday      9  2014     0       6        Monday_0h  \n",
       "4         Monday      9  2014     0      11        Monday_0h  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "827c9cac-77aa-47bd-b486-46ffee3bf9a9",
   "metadata": {},
   "outputs": [],
   "source": [
    "numeric_features = ['Lat', 'Lon']\n",
    "#categorical_features = ['Base']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "96c4bc44-340f-455c-9489-12015f7b3510",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create pipeline for numeric features\n",
    "numeric_transformer = Pipeline(steps=[\n",
    "    #('imputer', SimpleImputer()), \n",
    "    ('scaler', StandardScaler())\n",
    "])\n",
    "\n",
    "# Create pipeline for categorical features\n",
    "categorical_transformer = Pipeline(\n",
    "    steps=[\n",
    "    ('imputer', SimpleImputer(strategy='most_frequent')), # missing values will be replaced by most frequent value\n",
    "    ('encoder', OneHotEncoder(drop='first')) # first column will be dropped to avoid creating correlations between features\n",
    "    ])\n",
    "\n",
    "# Use ColumnTransformer to make a preprocessor object that describes all the treatments to be done\n",
    "preprocessor = ColumnTransformer(\n",
    "    transformers=[\n",
    "        ('num', numeric_transformer, numeric_features),\n",
    "        #('cat', categorical_transformer, categorical_features)\n",
    "    ])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "f480763e-8096-4716-9845-2a1289e352d0",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = preprocessor.fit_transform(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "b5eff682",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-12.71465439,  -0.51931048],\n",
       "       [  0.26399745,  -0.52959958],\n",
       "       [  0.40850405,  -0.25007886],\n",
       "       ...,\n",
       "       [  2.75979784,   0.68451499],\n",
       "       [ -0.76224431,  -0.59647878],\n",
       "       [ -0.61773772,   0.38098635]])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6156c8f1-543d-44b3-a714-6bdebbb73943",
   "metadata": {},
   "source": [
    "## EDA"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "d44e2ae9",
   "metadata": {},
   "source": [
    "#### Journeys By Month"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "5bb6ff2b",
   "metadata": {},
   "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/Time</th>\n",
       "      <th>nb_journeys</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2014-09-01 00:00:00</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2014-09-01 00:01:00</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2014-09-01 00:02:00</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2014-09-01 00:03:00</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2014-09-01 00:04:00</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Date/Time  nb_journeys\n",
       "0 2014-09-01 00:00:00           13\n",
       "1 2014-09-01 00:01:00           18\n",
       "2 2014-09-01 00:02:00           13\n",
       "3 2014-09-01 00:03:00           10\n",
       "4 2014-09-01 00:04:00           14"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pickup_time_df = df.groupby(by = \"Date/Time\").size().to_frame().reset_index()\n",
    "pickup_time_df.rename(columns = {0:'nb_journeys'}, inplace = True)\n",
    "pickup_time_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4a6ea9c1",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Visualising number of journeys depending on the day of the month and the hour of the day.\n",
    "fig = px.line(x=pickup_time_df[\"Date/Time\"], y = pickup_time_df[\"nb_journeys\"])\n",
    "fig.update_layout(\n",
    "    title=\"Evolution of number of Uber journeys over a month\",\n",
    "    xaxis_title=\"Time\",\n",
    "    yaxis_title=\"Nb journeys\"\n",
    "    )\n",
    "fig.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "deb43430",
   "metadata": {},
   "source": [
    "#### Journeys By Month Day"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "b6699ef2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1080x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "uber_monthdays = df.pivot_table(index=['Day'],\n",
    "                                  values='Base',\n",
    "                                  aggfunc='count')\n",
    "uber_monthdays.plot(kind='bar', figsize=(15,8))\n",
    "plt.ylabel('Total Journeys')\n",
    "plt.title('Journeys by Month Day');"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "f088fca5-f10e-42db-9bbd-b4a95941a144",
   "metadata": {},
   "source": [
    "#### Journeys By Weakdays"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "07dec487-f63d-4158-b868-2ed2078aad9d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:title={'center':'Bar Plot for Weekdays trips'}, xlabel='day'>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "uber_weekdays = df['DayOfWeekLabel'].value_counts().reset_index().rename(columns={'index':'day','DayOfWeekLabel':'count'})\n",
    "\n",
    "uber_weekdays.plot(\n",
    "    kind='bar',\n",
    "    x='day',\n",
    "    y='count',\n",
    "    title='Bar Plot for Weekdays trips',\n",
    "    figsize=(10, 5)\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0b299ad8",
   "metadata": {},
   "source": [
    "#### Journeys by Hour"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "13464a6a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1080x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "uber_hour = df.pivot_table(index=['Hour'],\n",
    "                                  values='Base',\n",
    "                                  aggfunc='count')\n",
    "uber_hour.plot(kind='bar', figsize=(15,8))\n",
    "plt.ylabel('Total Journeys')\n",
    "plt.title('Journeys by Hour');"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6e84bdfe-195c-49c8-be5c-47a8d7deae04",
   "metadata": {},
   "source": [
    "## Model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "40c80f25",
   "metadata": {},
   "source": [
    "##### Select $K$ optimal clusters"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2db13dcf",
   "metadata": {},
   "source": [
    "##### Elbow Method"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b1285b17",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Let's create a loop that will collect the Within-sum-of-square (wcss) for each value K \n",
    "# Let's use .inertia_ parameter to get the within sum of square value for each value K \n",
    "wcss =  []\n",
    "k = []\n",
    "for i in range (2,11): \n",
    "    kmeans = KMeans(n_clusters= i, random_state = 0)\n",
    "    kmeans.fit(X)\n",
    "    wcss.append(kmeans.inertia_)\n",
    "    k.append(i)\n",
    "    # print(\"WCSS for K={} --> {}\".format(i, wcss[-1]))\n",
    "\n",
    "print(wcss)\n",
    "\n",
    "# Create DataFrame\n",
    "wcss_frame = pd.DataFrame(wcss)\n",
    "k_frame = pd.Series(k)\n",
    "\n",
    "# Create figure\n",
    "fig= px.line(\n",
    "    wcss_frame,\n",
    "    x=k_frame,\n",
    "    y=wcss_frame.iloc[:,-1]\n",
    ")\n",
    "\n",
    "# Create title and axis labels\n",
    "fig.update_layout(\n",
    "    yaxis_title=\"Inertia\",\n",
    "    xaxis_title=\"# Clusters\",\n",
    "    title=\"Inertia per cluster\"\n",
    ")\n",
    "\n",
    "# Render\n",
    "#fig.show(renderer=\"notebook\")\n",
    "# fig.show(renderer=\"iframe\") # if using workspace\n",
    "fig.show() # if using workspace"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "56aabae1",
   "metadata": {},
   "source": [
    "## Silhouette 👥\n",
    "\n",
    "Let's now double check with the Silhouette method. We will use Sklearn again to get <a href=\"https://scikit-learn.org/stable/modules/generated/sklearn.metrics.silhouette_score.html?highlight=silhouette#sklearn.metrics.silhouette_score\" target=\"_blank\">silhouette scores.</a>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "da4fbd01",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import silhouette score\n",
    "from sklearn.metrics import silhouette_score\n",
    "\n",
    "# Computer mean silhouette score\n",
    "sil = []\n",
    "k = []\n",
    "\n",
    "## Careful, you need to start at i=2 as silhouette score cannot accept less than 2 labels \n",
    "for i in range (2,11): \n",
    "    kmeans = KMeans(n_clusters= i, random_state = 0)\n",
    "    kmeans.fit(X)\n",
    "    sil.append(silhouette_score(X, kmeans.predict(X)))\n",
    "    k.append(i)\n",
    "    #print(\"Silhouette score for K={} is {}\".format(i, sil[-1]))\n",
    "\n",
    "print(sil)\n",
    "\n",
    "# Create a data frame \n",
    "cluster_scores=pd.DataFrame(sil)\n",
    "k_frame = pd.Series(k)\n",
    "\n",
    "# Create figure\n",
    "fig = px.bar(data_frame=cluster_scores,  \n",
    "             x=k, \n",
    "             y=cluster_scores.iloc[:, -1]\n",
    "            )\n",
    "\n",
    "# Add title and axis labels\n",
    "fig.update_layout(\n",
    "    yaxis_title=\"Silhouette Score\",\n",
    "    xaxis_title=\"# Clusters\",\n",
    "    title=\"Silhouette Score per cluster\"\n",
    ")\n",
    "\n",
    "# Render\n",
    "#fig.show(renderer=\"notebook\")\n",
    "# fig.show(renderer=\"iframe\") # if using workspace\n",
    "fig.show() # if using workspace"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "589db10a-6817-4a90-8c7d-f9a44678ba85",
   "metadata": {},
   "outputs": [],
   "source": [
    "# define the model\n",
    "model = KMeans(n_clusters=4)\n",
    "\n",
    "# fit the model\n",
    "model.fit(X)\n",
    "\n",
    "# the prediction assign each example (pickup) to his cluster\n",
    "clusters = model.predict(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "0934e189",
   "metadata": {},
   "outputs": [],
   "source": [
    "df['Cluster'] = clusters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "31521bee",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Visualising the clusters on the map\n",
    "fig = px.scatter_mapbox(df,\n",
    "                        lat='Lat', \n",
    "                        lon='Lon',\n",
    "                        color='Cluster',\n",
    "                        zoom=5,\n",
    "                        height=800,\n",
    "                        mapbox_style='carto-positron'\n",
    "                        )\n",
    "fig.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "393477a3",
   "metadata": {},
   "source": [
    "> #### I have my <font color='green'>4 clusters</font>, Go ahead and Analyze those clusters."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e9e9a03c-df56-45dd-b9e7-6d990f75c6f6",
   "metadata": {},
   "source": [
    "## Analysing the results of K-means Clustering"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "75decdc2",
   "metadata": {},
   "source": [
    "> #### With so many data points, it's a difficult to pick out the characteristics of the clusters."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "6dc6136a",
   "metadata": {},
   "source": [
    "> #### But Centroid (a new datapoint representing the center of each cluster) is very useful to summary each cluster. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "b3d12ec4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 40.80616216, -73.89663687],\n",
       "       [ 40.70850756, -73.99155284],\n",
       "       [ 40.6694178 , -73.76938181],\n",
       "       [ 40.75690191, -73.98122129]])"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Bring the scaled data back to original values\n",
    "scaler = preprocessor.named_transformers_['num'].named_steps['scaler']\n",
    "X_centroids = scaler.inverse_transform(model.cluster_centers_)\n",
    "X_centroids"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "08120638",
   "metadata": {},
   "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>Lat</th>\n",
       "      <th>Lon</th>\n",
       "      <th>Cluster</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>40.806162</td>\n",
       "      <td>-73.896637</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>40.708508</td>\n",
       "      <td>-73.991553</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>40.669418</td>\n",
       "      <td>-73.769382</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>40.756902</td>\n",
       "      <td>-73.981221</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         Lat        Lon  Cluster\n",
       "0  40.806162 -73.896637        0\n",
       "1  40.708508 -73.991553        1\n",
       "2  40.669418 -73.769382        2\n",
       "3  40.756902 -73.981221        3"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "centroids = pd.DataFrame(X_centroids, columns=['Lat', 'Lon'])\n",
    "centroids['Cluster'] = centroids.index\n",
    "centroids"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f1de0647",
   "metadata": {},
   "outputs": [],
   "source": [
    "fig = px.scatter_mapbox(centroids,\n",
    "                        lat='Lat', \n",
    "                        lon='Lon',\n",
    "                        color='Cluster',\n",
    "                        hover_data=['Lat', 'Lon'],\n",
    "                        zoom=10,\n",
    "                        size=np.full(centroids.shape[0], 3),\n",
    "                        height=800,\n",
    "                        center=dict(lat=40.73, lon=-73.93),\n",
    "                        mapbox_style='carto-positron',\n",
    "                        title='Cluster Centroids'\n",
    "                        )\n",
    "fig.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "0fa5e92b",
   "metadata": {},
   "source": [
    "## The Goal is to visualize hot-zones per hour and per day of week"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "47ae1a80",
   "metadata": {},
   "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>DayOfWeekAndHour</th>\n",
       "      <th>Cluster</th>\n",
       "      <th>nb_journeys</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>667</th>\n",
       "      <td>Friday_3h</td>\n",
       "      <td>2</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>668</th>\n",
       "      <td>Saturday_1h</td>\n",
       "      <td>2</td>\n",
       "      <td>56</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>669</th>\n",
       "      <td>Sunday_2h</td>\n",
       "      <td>2</td>\n",
       "      <td>44</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>670</th>\n",
       "      <td>Sunday_7h</td>\n",
       "      <td>2</td>\n",
       "      <td>69</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>671</th>\n",
       "      <td>Thursday_2h</td>\n",
       "      <td>2</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    DayOfWeekAndHour  Cluster  nb_journeys\n",
       "667        Friday_3h        2           25\n",
       "668      Saturday_1h        2           56\n",
       "669        Sunday_2h        2           44\n",
       "670        Sunday_7h        2           69\n",
       "671      Thursday_2h        2           10"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Determining the number of pickup points for each hour of the week\n",
    "nb_journeys_df = df.groupby(['DayOfWeekAndHour', 'Cluster'], sort=False).size().reset_index()\n",
    "nb_journeys_df.rename(columns ={0: 'nb_journeys'}, inplace=True)\n",
    "nb_journeys_df.tail(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "4d2ac228",
   "metadata": {},
   "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>DayOfWeekAndHour</th>\n",
       "      <th>Cluster</th>\n",
       "      <th>centroid_latitude</th>\n",
       "      <th>centroid_longitude</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Monday_0h</td>\n",
       "      <td>2</td>\n",
       "      <td>40.665187</td>\n",
       "      <td>-73.759697</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Monday_0h</td>\n",
       "      <td>3</td>\n",
       "      <td>40.753210</td>\n",
       "      <td>-73.985474</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Monday_0h</td>\n",
       "      <td>0</td>\n",
       "      <td>40.814671</td>\n",
       "      <td>-73.913115</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Monday_0h</td>\n",
       "      <td>1</td>\n",
       "      <td>40.705426</td>\n",
       "      <td>-73.984955</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Monday_1h</td>\n",
       "      <td>3</td>\n",
       "      <td>40.752382</td>\n",
       "      <td>-73.986121</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  DayOfWeekAndHour  Cluster  centroid_latitude  centroid_longitude\n",
       "0        Monday_0h        2          40.665187          -73.759697\n",
       "1        Monday_0h        3          40.753210          -73.985474\n",
       "2        Monday_0h        0          40.814671          -73.913115\n",
       "3        Monday_0h        1          40.705426          -73.984955\n",
       "4        Monday_1h        3          40.752382          -73.986121"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "centroid_coordinates =  df.groupby(['DayOfWeekAndHour', 'Cluster'], sort=False).apply(calculate_centroid).reset_index()\n",
    "centroid_coordinates.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "929ae5eb",
   "metadata": {},
   "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>DayOfWeekAndHour</th>\n",
       "      <th>Cluster</th>\n",
       "      <th>Lat</th>\n",
       "      <th>Lon</th>\n",
       "      <th>Day</th>\n",
       "      <th>DayOfWeek</th>\n",
       "      <th>Month</th>\n",
       "      <th>Year</th>\n",
       "      <th>Hour</th>\n",
       "      <th>Minute</th>\n",
       "      <th>nb_journeys</th>\n",
       "      <th>centroid_latitude</th>\n",
       "      <th>centroid_longitude</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Monday_0h</td>\n",
       "      <td>2</td>\n",
       "      <td>40.6483</td>\n",
       "      <td>-73.78260</td>\n",
       "      <td>22.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2014.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>130</td>\n",
       "      <td>40.665187</td>\n",
       "      <td>-73.759697</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Monday_0h</td>\n",
       "      <td>3</td>\n",
       "      <td>40.7503</td>\n",
       "      <td>-73.98720</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2014.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>707</td>\n",
       "      <td>40.753210</td>\n",
       "      <td>-73.985474</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Monday_0h</td>\n",
       "      <td>0</td>\n",
       "      <td>40.8087</td>\n",
       "      <td>-73.93530</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2014.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>28.0</td>\n",
       "      <td>137</td>\n",
       "      <td>40.814671</td>\n",
       "      <td>-73.913115</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Monday_0h</td>\n",
       "      <td>1</td>\n",
       "      <td>40.7148</td>\n",
       "      <td>-73.98730</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2014.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>28.0</td>\n",
       "      <td>911</td>\n",
       "      <td>40.705426</td>\n",
       "      <td>-73.984955</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Monday_1h</td>\n",
       "      <td>3</td>\n",
       "      <td>40.7508</td>\n",
       "      <td>-73.99015</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2014.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>490</td>\n",
       "      <td>40.752382</td>\n",
       "      <td>-73.986121</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  DayOfWeekAndHour  Cluster      Lat       Lon   Day  DayOfWeek  Month  \\\n",
       "0        Monday_0h        2  40.6483 -73.78260  22.0        0.0    9.0   \n",
       "1        Monday_0h        3  40.7503 -73.98720   8.0        0.0    9.0   \n",
       "2        Monday_0h        0  40.8087 -73.93530   8.0        0.0    9.0   \n",
       "3        Monday_0h        1  40.7148 -73.98730   8.0        0.0    9.0   \n",
       "4        Monday_1h        3  40.7508 -73.99015   8.0        0.0    9.0   \n",
       "\n",
       "     Year  Hour  Minute  nb_journeys  centroid_latitude  centroid_longitude  \n",
       "0  2014.0   0.0    15.0          130          40.665187          -73.759697  \n",
       "1  2014.0   0.0    27.0          707          40.753210          -73.985474  \n",
       "2  2014.0   0.0    28.0          137          40.814671          -73.913115  \n",
       "3  2014.0   0.0    28.0          911          40.705426          -73.984955  \n",
       "4  2014.0   1.0    27.0          490          40.752382          -73.986121  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Nb of rows in the dataframe: 672\n"
     ]
    }
   ],
   "source": [
    "# Storing journey numbers as a pandas series\n",
    "nb_journeys = nb_journeys_df['nb_journeys']\n",
    "\n",
    "# Finding the optimal driver position for each cluster by taking the centroid of the coordinates of the pickup points included in this cluster. \n",
    "# the centroid is more better than the median or mean of the coordinates of the pickup points included in this cluster\n",
    "hotzones_df = df.groupby(['DayOfWeekAndHour', 'Cluster'], sort=False).median(numeric_only=True).reset_index()\n",
    "hotzones_df['nb_journeys'] = nb_journeys\n",
    "hotzones_df['centroid_latitude'] = centroid_coordinates['centroid_latitude']\n",
    "hotzones_df['centroid_longitude'] = centroid_coordinates['centroid_longitude']\n",
    "\n",
    "# Checking the resulting dataframe\n",
    "display(hotzones_df.head(5))\n",
    "\n",
    "# Checking the length of the dataframe\n",
    "print('Nb of rows in the dataframe:', len(hotzones_df))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3ed03741",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Lets us look at an animated map where each frame shows pickup points in a given hour of a day\n",
    "fig = px.scatter_mapbox(hotzones_df, \n",
    "                        lat=\"centroid_latitude\",\n",
    "                        lon=\"centroid_longitude\",\n",
    "                        animation_frame='DayOfWeekAndHour',\n",
    "                        zoom=10,\n",
    "                        color=\"Cluster\",\n",
    "                        size=np.full(hotzones_df.shape[0], 0.1),\n",
    "                        mapbox_style='carto-positron',\n",
    "                        height=800,\n",
    "                        title='Clusters Recommended positions per Hour and per Day of week'\n",
    "                        )\n",
    "fig.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3ae75bef",
   "metadata": {},
   "source": [
    "## Comparaison with DBSCAN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "addb5c62",
   "metadata": {},
   "outputs": [],
   "source": [
    "size = 1000\n",
    "df_new = df.copy().head(size)\n",
    "X = preprocessor.fit_transform(df_new)\n",
    "\n",
    "# Define the distance metric to use (in this case, Haversine)\n",
    "# The Haversine distance is commonly used for calculating distances between points on the surface of the earth\n",
    "dist = DistanceMetric.get_metric('haversine')\n",
    "\n",
    "# Calculate the distance matrix\n",
    "dist_matrix = dist.pairwise(np.radians(X))\n",
    "\n",
    "# Apply DBSCAN clustering\n",
    "dbscan = DBSCAN(eps=0.1, min_samples=10, metric='precomputed')\n",
    "dbscan.fit(dist_matrix)\n",
    "\n",
    "# Get the cluster labels and the number of clusters\n",
    "labels = dbscan.labels_\n",
    "n_clusters = len(set(labels)) - (1 if -1 in labels else 0)\n",
    "\n",
    "print(\"Estimated number of clusters: %d\" % n_clusters)\n",
    "\n",
    "# Assigning a cluster to each row in the dataframe\n",
    "df_new[\"Cluster_DB\"] = labels\n",
    "\n",
    "df_new = df_new[df_new['Cluster_DB'] != -1]\n",
    "\n",
    "# Determining the number of pickup points for each hour of the week\n",
    "nb_journeys_df = df_new.groupby(['DayOfWeekAndHour', 'Cluster_DB'], sort=False).size().reset_index()\n",
    "nb_journeys_df.rename(columns ={0: 'nb_journeys'}, inplace=True)\n",
    "\n",
    "centroid_coordinates =  df_new.groupby(['DayOfWeekAndHour', 'Cluster_DB'], sort=False).apply(calculate_centroid).reset_index()\n",
    "centroid_coordinates.head()\n",
    "\n",
    "# Storing journey numbers as a pandas series\n",
    "nb_journeys = nb_journeys_df['nb_journeys']\n",
    "\n",
    "# Finding the optimal driver position for each cluster by taking the centroid of the coordinates of the pickup points included in this cluster. \n",
    "# the centroid is more better than the median or mean of the coordinates of the pickup points included in this cluster\n",
    "hotzones_df = df_new.groupby(['DayOfWeekAndHour', 'Cluster_DB'], sort=False).median(numeric_only=True).reset_index()\n",
    "hotzones_df['nb_journeys'] = nb_journeys\n",
    "hotzones_df['centroid_latitude'] = centroid_coordinates['centroid_latitude']\n",
    "hotzones_df['centroid_longitude'] = centroid_coordinates['centroid_longitude']\n",
    "\n",
    "# Lets us look at an animated map where each frame shows pickup points in a given hour of a day\n",
    "fig = px.scatter_mapbox(hotzones_df, \n",
    "                        lat=\"centroid_latitude\",\n",
    "                        lon=\"centroid_longitude\",\n",
    "                        animation_frame='DayOfWeekAndHour',\n",
    "                        zoom=10,\n",
    "                        color=\"Cluster_DB\",\n",
    "                        size=np.full(hotzones_df.shape[0], 0.1),\n",
    "                        mapbox_style='carto-positron',\n",
    "                        height=800,\n",
    "                        title='DBSCAN Clusters Recommended positions per Hour and per Day of week'\n",
    "                        )\n",
    "fig.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f18ad936",
   "metadata": {},
   "source": [
    "### Sources\n",
    "\n",
    "- [does-the-mean-of-latitude-and-longitude-make-sense](https://gis.stackexchange.com/questions/164322/does-the-mean-of-latitude-and-longitude-make-sense)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.12"
  },
  "vscode": {
   "interpreter": {
    "hash": "c766f7c526f554c39534ac5a6b6631b3579ceeff597e65c8c23f9bc665d77d04"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}