5. Moving Window Functions.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%matplotlib inline \n",
    "import matplotlib.pylab\n",
    "import numpy as np\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Window functions are like aggregation functions\n",
    "# You can use them in conjunction with .resample() which will soon be a deferred operation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "df = pd.DataFrame(np.random.randn(600, 3), index = pd.date_range('7/1/2016', freq = 'S', periods = 600), columns = ['A', 'B', 'C'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#pd.rolling_mean(df, window = 2)[1:10] # in future versions you want to resample separately\n",
    "r = df.rolling(window = 10)\n",
    "#r.agg, r.apply, r.count, r.exclusions, r.max, r.median, r.name, r.quantile, r.kurt, r.cov, r.corr, r.aggregate, r.std, r.skew, r.sum, r.var\n",
    "df.plot(style = 'k--')\n",
    "r.mean().plot(style = 'k')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#exponentially weighted\n",
    "pd.ewma(df, span = 15, min_periods = 2)[1:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# can also easily plot a rolling average per column\n",
    "df = pd.DataFrame(np.random.randn(1000, 4), index = pd.date_range('6/6/16', periods = 1000), columns = ['A', 'B', 'C', 'D'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "df = df.cumsum()\n",
    "df.rolling(window = 50).sum().plot(subplots=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# What about a custom function?\n",
    "df.rolling(window = 10).apply(lambda x: np.fabs(x - x.mean()).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Expanding windows\n",
    "# Yields the value of the statistic with all the data available up to that point in time\n",
    "df.expanding(min_periods = 1).mean()[1:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Will an expanding window provide a more or less stable statistic than a rolling statistic? \n",
    "# When might one vs. the other be appropriate?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# EWMA\n",
    "# Why use EWMA?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Go do it!\n",
    "# pandas.pydata.org/pandas-docs/stable/computation.html#stats-aggregate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Come up with a series that shows how EWMA and Rolling can importantly differ even with the same window setting"
   ]
  }
 ],
 "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": "3.5.1"
  },
  "widgets": {
   "state": {},
   "version": "1.1.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}