Statsmodels example: Using dates with timeseries models

ex_dates.ipynb

{
 "metadata": {
  "name": "ex_dates"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "# Using dates with timeseries models"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import statsmodels.api as sm\n",
      "import numpy as np\n",
      "import pandas"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 1
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "## Getting started"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "data = sm.datasets.sunspots.load()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 2
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Right now an annual date series must be datetimes at the end of the year."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from datetime import datetime\n",
      "dates = sm.tsa.datetools.dates_from_range('1700', length=len(data.endog))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 3
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "## Using Pandas\n",
      "\n",
      "Make a pandas TimeSeries or DataFrame"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "endog = pandas.TimeSeries(data.endog, index=dates)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 4
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Instantiate the model"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "ar_model = sm.tsa.AR(endog, freq='A')\n",
      "pandas_ar_res = ar_model.fit(maxlag=9, method='mle', disp=-1)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 5
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Out-of-sample prediction"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "pred = pandas_ar_res.predict(start='2005', end='2015')\n",
      "print pred"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "2005-12-31    20.003275\n",
        "2006-12-31    24.703970\n",
        "2007-12-31    20.026113\n",
        "2008-12-31    23.473658\n",
        "2009-12-31    30.858584\n",
        "2010-12-31    61.335478\n",
        "2011-12-31    87.024727\n",
        "2012-12-31    91.321290\n",
        "2013-12-31    79.921658\n",
        "2014-12-31    60.799537\n",
        "2015-12-31    40.374871\n",
        "Freq: A-DEC\n"
       ]
      }
     ],
     "prompt_number": 6
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "## Using explicit dates"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "ar_model = sm.tsa.AR(data.endog, dates=dates, freq='A')\n",
      "ar_res = ar_model.fit(maxlag=9, method='mle', disp=-1)\n",
      "pred = ar_res.predict(start='2005', end='2015')\n",
      "print pred"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[ 20.00327492  24.70396965  20.02611309  23.47365775  30.8585841\n",
        "  61.33547797  87.02472654  91.3212902   79.92165773  60.79953659\n",
        "  40.37487104]\n"
       ]
      }
     ],
     "prompt_number": 7
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "This just returns a regular array, but since the model has date information attached, you can get the prediction dates in a roundabout way."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print ar_res._data.predict_dates"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "<class 'pandas.tseries.index.DatetimeIndex'>\n",
        "[2005-12-31 00:00:00, ..., 2015-12-31 00:00:00]\n",
        "Length: 11, Freq: A-DEC, Timezone: None\n"
       ]
      }
     ],
     "prompt_number": 8
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Note: This attribute only exists if predict has been called. It holds the dates associated with the last call to predict."
     ]
    }
   ],
   "metadata": {}
  }
 ]
}