examples/Naive.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Naive predictions\n",
    "\n",
    "Two \"predictions\" which are very simple.  Surely any reasonable algorithm should be able to out-perform these.  These are not taken directly from the literature, but are \"obvious\" once you've spent some time with the ideas."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Naive (\"counting grid\") algorithm\n",
    "\n",
    "This is a grid based algorithm.  A grid is laid down on the study region.  We expect that small (50m or less) will perform badly, compared to larger grid sizes.\n",
    "\n",
    "### Aim of the algorithm\n",
    "\n",
    "We look at all data before the current time (or perhaps a \"sliding window\" of the last $n$ months before the current time).  For each grid cell independently, we count the number of events which have occurred in that cell.  This gives the (non-normalised) risk in that cell.\n",
    "\n",
    "Notice then that there is no direct consideration of time (unless a sliding window is used).  There is no consideration of space, so this algorithm will suffer from the [MAUP](https://en.wikipedia.org/wiki/Modifiable_areal_unit_problem) as well as taking no account of likely movement of offenders.\n",
    "\n",
    "### Prediction\n",
    "\n",
    "As usual, we can look at the top 1%, 5%, 10% of cells by risk, and so forth."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Naive KDE\n",
    "\n",
    "This algorithm is in some sense the most naive answer to the problem that the \"counting grid\" algorithm suffers from.  The output prediction is in the form of a grid, but this is computed only at the very end, so really the algorith is \"continuous\".\n",
    "\n",
    "We take an out the box [kernel density estimation](https://en.wikipedia.org/wiki/Kernel_density_estimation) algorithm, feed in the spatial locations of past crime events (so again, we use all crimes before the current time, and ignore timestamps) and then sample the resulting kernel to a grid.\n",
    "\n",
    "A KDE is an estimate of a probability density of the form\n",
    "\\[ f(x) = \\frac{1}{n} \\sum_{i=1}^n K(x - x_i) \\]\n",
    "where $(x_i)$ are the (two dimensional vector) locations of the events.  Here $K$ is the kernel, which is generality can be any two-dimensional function which is positive, and has $\\int_{\\mathbb R^2} K = 1$.\n",
    "\n",
    "We use the `scipy` Gaussian KDE.  This first computes the [sample covariance](https://en.wikipedia.org/wiki/Sample_mean_and_covariance) of the input data, and then transforms the data using this to have unit covariance.  Then a \"plug in bandwidth rule\" is used to choose a value of $h>0$.  Finally $K$ is equal to\n",
    "\\[ K(x) = \\frac{1}{2\\pi h^2} \\exp(- \\|x\\|^2 / 2h^2). \\]\n",
    "For more details, see the `kernel_estimation.ipynb` notebook under `notebooks`.\n",
    "\n",
    "A more sophisticated KDE would take account of edge effects (we typically only look at a finite area, while the KDE $f$ will have infinite support, so \"predicting\" some events outside of the study area).\n",
    "\n",
    "This tends to give a very (over) smoothly varying kernel."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import open_cp\n",
    "import open_cp.naive as naive"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Generate some random data\n",
    "import datetime\n",
    "size = 30\n",
    "times = [datetime.datetime(2017,3,10) + datetime.timedelta(days=np.random.randint(0,10)) for _ in range(size)]\n",
    "times.sort()\n",
    "xc = np.random.random(size=size) * 500\n",
    "yc = np.random.random(size=size) * 500\n",
    "points = open_cp.TimedPoints.from_coords(times, xc, yc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridPredictionArray(offset=(0,0), size=50x50, risk intensity size=10x10)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "region = open_cp.RectangularRegion(0,500, 0,500)\n",
    "predictor = naive.CountingGridKernel(50, region=region)\n",
    "predictor.data = points\n",
    "prediction = predictor.predict()\n",
    "prediction"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The following is the \"counting grid\" predictor.  Grid cells which contain 2 past events are ranked higher in risk than those which contain only 1 past event."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkkAAAJDCAYAAAAfG4BSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fea138f4630>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(10,10))\n",
    "\n",
    "m = ax.pcolor(*prediction.mesh_data(), prediction.intensity_matrix)\n",
    "ax.scatter(points.xcoords, points.ycoords, marker=\"+\", color=\"black\")\n",
    "ax.set(xlim=[0, 500], ylim=[0, 500])\n",
    "cb = plt.colorbar(m, ax=ax)\n",
    "cb.set_label(\"Relative risk\")\n",
    "None"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We now use the `scipy` Gaussian KDE.  We produce a \"continuous\" prediction, and then sample to a grid of the same size and location as that used before.  Notice that the result is a lot \"smoother\"."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<open_cp.predictors.KernelRiskPredictor at 0x7fea1a002710>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictor = naive.ScipyKDE()\n",
    "predictor.data = points\n",
    "prediction = predictor.predict()\n",
    "prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridPredictionArray(offset=(0,0), size=50x50, risk intensity size=10x10)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gridpred = open_cp.predictors.GridPredictionArray.from_continuous_prediction_region(prediction, region, 50)\n",
    "gridpred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmIAAAJDCAYAAACopnh1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fea10e3fb70>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(10,10))\n",
    "\n",
    "m = ax.pcolor(*gridpred.mesh_data(), gridpred.intensity_matrix)\n",
    "ax.scatter(points.xcoords, points.ycoords, marker=\"+\", color=\"black\")\n",
    "ax.set(xlim=[0, 500], ylim=[0, 500])\n",
    "cb = plt.colorbar(m, ax=ax)\n",
    "cb.set_label(\"Relative risk\")\n",
    "None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "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": "3.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}