notebook/ch10_working_with_data.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Populating the interactive namespace from numpy and matplotlib\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: pylab import has clobbered these variables: ['shape', 'random', 'dot', 'mean']\n",
      "`%matplotlib` prevents importing * from pylab and numpy\n"
     ]
    }
   ],
   "source": [
    "from collections import Counter, defaultdict\n",
    "from functools import partial, reduce\n",
    "from linear_algebra import shape, get_row, get_column, make_matrix, \\\n",
    "    vector_mean, vector_sum, dot, magnitude, vector_subtract, scalar_multiply\n",
    "from stats import correlation, standard_deviation, mean\n",
    "from probability import inverse_normal_cdf\n",
    "from gradient_descent import maximize_batch\n",
    "import math, random, csv\n",
    "import matplotlib.pyplot as plt\n",
    "import dateutil.parser\n",
    "\n",
    "%pylab inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def bucketize(point, bucket_size):\n",
    "    \"\"\"floor the point to the next lower multiple of bucket_size\"\"\"\n",
    "    return bucket_size * math.floor(point / bucket_size)\n",
    "\n",
    "def make_histogram(points, bucket_size):\n",
    "    \"\"\"buckets the points and counts how many in each bucket\"\"\"\n",
    "    return Counter(bucketize(point, bucket_size) for point in points)\n",
    "\n",
    "def plot_histogram(points, bucket_size, title=\"\"):\n",
    "    histogram = make_histogram(points, bucket_size)\n",
    "    plt.bar(histogram.keys(), histogram.values(), width=bucket_size)\n",
    "    plt.title(title)\n",
    "    plt.show()\n",
    "\n",
    "def compare_two_distributions():\n",
    "\n",
    "    random.seed(0)\n",
    "\n",
    "    uniform = [random.randint(-100,101) for _ in range(200)]\n",
    "    normal = [57 * inverse_normal_cdf(random.random())\n",
    "              for _ in range(200)]\n",
    "\n",
    "    plot_histogram(uniform, 10, \"Uniform Histogram\")\n",
    "    plot_histogram(normal, 10, \"Normal Histogram\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEKCAYAAADpfBXhAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10f5b85f8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEKCAYAAADpfBXhAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10f6d2b00>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "compare_two_distributions()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def random_normal():\n",
    "    \"\"\"returns a random draw from a standard normal distribution\"\"\"\n",
    "    return inverse_normal_cdf(random.random())\n",
    "\n",
    "xs = [random_normal() for _ in range(1000)]\n",
    "ys1 = [ x + random_normal() / 2 for x in xs]\n",
    "ys2 = [-x + random_normal() / 2 for x in xs]\n",
    "\n",
    "\n",
    "def scatter():\n",
    "    plt.scatter(xs, ys1, marker='.', color='black', label='ys1')\n",
    "    plt.scatter(xs, ys2, marker='.', color='gray',  label='ys2')\n",
    "    plt.xlabel('xs')\n",
    "    plt.ylabel('ys')\n",
    "    plt.legend(loc=9)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEPCAYAAABP1MOPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11256d198>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "scatter()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8973760231027484"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "correlation(xs,ys1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-0.9033393123522656"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "correlation(xs,ys2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def correlation_matrix(data):\n",
    "    \"\"\"returns the num_columns x num_columns matrix whose (i, j)th entry\n",
    "    is the correlation between columns i and j of data\"\"\"\n",
    "\n",
    "    _, num_columns = shape(data)\n",
    "\n",
    "    def matrix_entry(i, j):\n",
    "        return correlation(get_column(data, i), get_column(data, j))\n",
    "\n",
    "    return make_matrix(num_columns, num_columns, matrix_entry)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def make_scatterplot_matrix():\n",
    "\n",
    "    # first, generate some random data\n",
    "\n",
    "    num_points = 100\n",
    "\n",
    "    def random_row():\n",
    "        row = [None, None, None, None]\n",
    "        row[0] = random_normal()\n",
    "        row[1] = -5 * row[0] + random_normal()\n",
    "        row[2] = row[0] + row[1] + 5 * random_normal()\n",
    "        row[3] = 6 if row[2] > -2 else 0\n",
    "        return row\n",
    "    random.seed(0)\n",
    "    data = [random_row()\n",
    "            for _ in range(num_points)]\n",
    "\n",
    "    # then plot it\n",
    "\n",
    "    _, num_columns = shape(data)\n",
    "    fig, ax = plt.subplots(num_columns, num_columns)\n",
    "\n",
    "    for i in range(num_columns):\n",
    "        for j in range(num_columns):\n",
    "\n",
    "            # scatter column_j on the x-axis vs column_i on the y-axis\n",
    "            if i != j: ax[i][j].scatter(get_column(data, j), get_column(data, i))\n",
    "\n",
    "            # unless i == j, in which case show the series name\n",
    "            else: ax[i][j].annotate(\"series \" + str(i), (0.5, 0.5),\n",
    "                                    xycoords='axes fraction',\n",
    "                                    ha=\"center\", va=\"center\")\n",
    "\n",
    "            # then hide axis labels except left and bottom charts\n",
    "            if i < num_columns - 1: ax[i][j].xaxis.set_visible(False)\n",
    "            if j > 0: ax[i][j].yaxis.set_visible(False)\n",
    "\n",
    "    # fix the bottom right and top left axis labels, which are wrong because\n",
    "    # their charts only have text in them\n",
    "    ax[-1][-1].set_xlim(ax[0][-1].get_xlim())\n",
    "    ax[0][0].set_ylim(ax[0][1].get_ylim())\n",
    "\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXkAAAEACAYAAABWLgY0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nGAz9GDPIGwwGQz/GDPIGg8HQj/l/4z2UUCxDabMAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1125bd7b8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "make_scatterplot_matrix()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def parse_row(input_row, parsers):\n",
    "    \"\"\"given a list of parsers (some of which may be None)\n",
    "    apply the appropriate one to each element of the input_row\"\"\"\n",
    "    return [parser(value) if parser is not None else value\n",
    "            for value, parser in zip(input_row, parsers)]\n",
    "\n",
    "def parse_rows_with(reader, parsers):\n",
    "    \"\"\"wrap a reader to apply the parsers to each of its rows\"\"\"\n",
    "    for row in reader:\n",
    "        yield parse_row(row, parsers)\n",
    "\n",
    "def try_or_none(f):\n",
    "    \"\"\"wraps f to return None if f raises an exception\n",
    "    assumes f takes only one input\"\"\"\n",
    "    def f_or_none(x):\n",
    "        try: return f(x)\n",
    "        except: return None\n",
    "    return f_or_none\n",
    "\n",
    "def parse_row(input_row, parsers):\n",
    "    return [try_or_none(parser)(value) if parser is not None else value\n",
    "            for value, parser in zip(input_row, parsers)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[datetime.datetime(2014, 6, 20, 0, 0), 'AAPL', 90.91]\n",
      "[datetime.datetime(2014, 6, 20, 0, 0), 'MSFT', 41.68]\n",
      "[datetime.datetime(3014, 6, 20, 0, 0), 'FB', 64.5]\n",
      "[datetime.datetime(2014, 6, 19, 0, 0), 'AAPL', 91.86]\n",
      "[datetime.datetime(2014, 6, 19, 0, 0), 'MSFT', None]\n",
      "[datetime.datetime(2014, 6, 19, 0, 0), 'FB', 64.34]\n"
     ]
    }
   ],
   "source": [
    "data = []\n",
    "\n",
    "with open(\"comma_delimited_stock_prices.csv\", \"r\", encoding='utf8', newline='') as f:\n",
    "    reader = csv.reader(f)\n",
    "    for line in parse_rows_with(reader, [dateutil.parser.parse, None, float]):\n",
    "        data.append(line)\n",
    "\n",
    "for row in data:\n",
    "    if any(x is None for x in row):\n",
    "        print(row)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def try_parse_field(field_name, value, parser_dict):\n",
    "    \"\"\"try to parse value using the appropriate function from parser_dict\"\"\"\n",
    "    parser = parser_dict.get(field_name) # None if no such entry\n",
    "    if parser is not None:\n",
    "        return try_or_none(parser)(value)\n",
    "    else:\n",
    "        return value\n",
    "    \n",
    "def parse_dict(input_dict, parser_dict):\n",
    "    return { field_name : try_parse_field(field_name, value, parser_dict)\n",
    "             for field_name, value in input_dict.items() }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "stocks\n",
      "max aapl price 119.0\n"
     ]
    }
   ],
   "source": [
    "print(\"stocks\")\n",
    "with open(\"stocks.txt\", \"r\", encoding='utf8', newline='') as f:\n",
    "    reader = csv.DictReader(f, delimiter=\"\\t\")\n",
    "    data = [parse_dict(row, { 'date' : dateutil.parser.parse,\n",
    "                                  'closing_price' : float })\n",
    "                for row in reader]\n",
    "\n",
    "max_aapl_price = max(row[\"closing_price\"]\n",
    "                         for row in data\n",
    "                         if row[\"symbol\"] == \"AAPL\")\n",
    "print(\"max aapl price\", max_aapl_price)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max price by symbol\n",
      "{'AAPL': 119.0, 'MSFT': 49.3, 'FB': 81.45}\n"
     ]
    }
   ],
   "source": [
    "# group rows by symbol\n",
    "by_symbol = defaultdict(list)\n",
    "\n",
    "for row in data:\n",
    "    by_symbol[row[\"symbol\"]].append(row)\n",
    "\n",
    "# use a dict comprehension to find the max for each symbol\n",
    "max_price_by_symbol = { symbol : max(row[\"closing_price\"]\n",
    "                            for row in grouped_rows)\n",
    "                            for symbol, grouped_rows in by_symbol.items() }\n",
    "print(\"max price by symbol\")\n",
    "print(max_price_by_symbol)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def picker(field_name):\n",
    "    \"\"\"returns a function that picks a field out of a dict\"\"\"\n",
    "    return lambda row: row[field_name]\n",
    "\n",
    "def pluck(field_name, rows):\n",
    "    \"\"\"turn a list of dicts into the list of field_name values\"\"\"\n",
    "    return map(picker(field_name), rows)\n",
    "\n",
    "def group_by(grouper, rows, value_transform=None):\n",
    "    # key is output of grouper, value is list of rows\n",
    "    grouped = defaultdict(list)\n",
    "    for row in rows:\n",
    "        grouped[grouper(row)].append(row)\n",
    "    if value_transform is None:\n",
    "        return grouped\n",
    "    else:\n",
    "        return { key : value_transform(rows)\n",
    "                 for key, rows in grouped.items() }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'AAPL': 119.0, 'FB': 81.45, 'MSFT': 49.3}"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "max_price_by_symbol = group_by(picker(\"symbol\"), data,\n",
    "                              lambda rows: max(pluck(\"closing_price\", rows) ))\n",
    "\n",
    "max_price_by_symbol"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def percent_price_change(yesterday, today):\n",
    "    return today[\"closing_price\"] / yesterday[\"closing_price\"] - 1\n",
    "\n",
    "def day_over_day_changes(grouped_rows):\n",
    "    # sort the rows by date\n",
    "    ordered = sorted(grouped_rows, key=picker(\"date\"))\n",
    "    # zip with an offset to get pairs of consecutive days\n",
    "    return [{ \"symbol\" : today[\"symbol\"],\n",
    "              \"date\" : today[\"date\"],\n",
    "              \"change\" : percent_price_change(yesterday, today) }\n",
    "             for yesterday, today in zip(ordered, ordered[1:])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max change {'change': 0.3283582089552237, 'date': datetime.datetime(1997, 8, 6, 0, 0), 'symbol': 'AAPL'}\n",
      "min change {'change': -0.5193370165745856, 'date': datetime.datetime(2000, 9, 29, 0, 0), 'symbol': 'AAPL'}\n"
     ]
    }
   ],
   "source": [
    "# key is symbol, value is list of \"change\" dicts\n",
    "changes_by_symbol = group_by(picker(\"symbol\"), data, day_over_day_changes)\n",
    "# collect all \"change\" dicts into one big list\n",
    "all_changes = [change\n",
    "                   for changes in changes_by_symbol.values()\n",
    "                   for change in changes]\n",
    "\n",
    "print(\"max change\", max(all_changes, key=picker(\"change\")))\n",
    "print(\"min change\", min(all_changes, key=picker(\"change\")))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "overall change by month\n",
      "{1: 19.972214514609085, 2: 0.11858483359215, 3: 2.818850048603327, 4: 6.9341967849847626, 5: 1.2555660890321603, 6: -0.5977889232201744, 7: -0.17183091713561482, 8: 6.220055959879489, 9: -0.741162543642831, 10: 21.552541251680015, 11: 3.758572336791725, 12: 2.2895332950255933}\n"
     ]
    }
   ],
   "source": [
    "# to combine percent changes, we add 1 to each, multiply them, and subtract 1\n",
    "# for instance, if we combine +10% and -20%, the overall change is\n",
    "# (1 + 10%) * (1 - 20%) - 1 = 1.1 * .8 - 1 = -12%\n",
    "def combine_pct_changes(pct_change1, pct_change2):\n",
    "    return (1 + pct_change1) * (1 + pct_change2) - 1\n",
    "\n",
    "def overall_change(changes):\n",
    "    return reduce(combine_pct_changes, pluck(\"change\", changes))\n",
    "\n",
    "overall_change_by_month = group_by(lambda row: row['date'].month,\n",
    "                                       all_changes,\n",
    "                                       overall_change)\n",
    "print(\"overall change by month\")\n",
    "print(overall_change_by_month)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def scale(data_matrix):\n",
    "    num_rows, num_cols = shape(data_matrix)\n",
    "    means = [mean(get_column(data_matrix,j))\n",
    "             for j in range(num_cols)]\n",
    "    stdevs = [standard_deviation(get_column(data_matrix,j))\n",
    "              for j in range(num_cols)]\n",
    "    return means, stdevs\n",
    "\n",
    "def rescale(data_matrix):\n",
    "    \"\"\"rescales the input data so that each column\n",
    "    has mean 0 and standard deviation 1\n",
    "    ignores columns with no deviation\"\"\"\n",
    "    means, stdevs = scale(data_matrix)\n",
    "\n",
    "    def rescaled(i, j):\n",
    "        if stdevs[j] > 0:\n",
    "            return (data_matrix[i][j] - means[j]) / stdevs[j]\n",
    "        else:\n",
    "            return data_matrix[i][j]\n",
    "\n",
    "    num_rows, num_cols = shape(data_matrix)\n",
    "    return make_matrix(num_rows, num_cols, rescaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rescaling\n",
      "original:  [[1, 20, 2], [1, 30, 3], [1, 40, 4]]\n",
      "scale:  ([1.0, 30.0, 3.0], [0.0, 10.0, 1.0])\n",
      "rescaled:  [[1, -1.0, -1.0], [1, 0.0, 0.0], [1, 1.0, 1.0]]\n"
     ]
    }
   ],
   "source": [
    "print(\"rescaling\")\n",
    "\n",
    "data = [[1, 20, 2],\n",
    "            [1, 30, 3],\n",
    "            [1, 40, 4]]\n",
    "\n",
    "print(\"original: \", data)\n",
    "print(\"scale: \", scale(data))\n",
    "print(\"rescaled: \", rescale(data))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXwAAAEACAYAAACwB81wAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x112c6b390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "X = [\n",
    "    [20.9666776351559,-13.1138080189357],\n",
    "    [22.7719907680008,-19.8890894944696],\n",
    "    [25.6687103160153,-11.9956004517219],\n",
    "    [18.0019794950564,-18.1989191165133],\n",
    "    [21.3967402102156,-10.8893126308196],\n",
    "    [0.443696899177716,-19.7221132386308],\n",
    "    [29.9198322142127,-14.0958668502427],\n",
    "    [19.0805843080126,-13.7888747608312],\n",
    "    [16.4685063521314,-11.2612927034291],\n",
    "    [21.4597664701884,-12.4740034586705],\n",
    "    [3.87655283720532,-17.575162461771],\n",
    "    [34.5713920556787,-10.705185165378],\n",
    "    [13.3732115747722,-16.7270274494424],\n",
    "    [20.7281704141919,-8.81165591556553],\n",
    "    [24.839851437942,-12.1240962157419],\n",
    "    [20.3019544741252,-12.8725060780898],\n",
    "    [21.9021426929599,-17.3225432396452],\n",
    "    [23.2285885715486,-12.2676568419045],\n",
    "    [28.5749111681851,-13.2616470619453],\n",
    "    [29.2957424128701,-14.6299928678996],\n",
    "    [15.2495527798625,-18.4649714274207],\n",
    "    [26.5567257400476,-9.19794350561966],\n",
    "    [30.1934232346361,-12.6272709845971],\n",
    "    [36.8267446011057,-7.25409849336718],\n",
    "    [32.157416823084,-10.4729534347553],\n",
    "    [5.85964365291694,-22.6573731626132],\n",
    "    [25.7426190674693,-14.8055803854566],\n",
    "    [16.237602636139,-16.5920595763719],\n",
    "    [14.7408608850568,-20.0537715298403],\n",
    "    [6.85907008242544,-18.3965586884781],\n",
    "    [26.5918329233128,-8.92664811750842],\n",
    "    [-11.2216019958228,-27.0519081982856],\n",
    "    [8.93593745011035,-20.8261235122575],\n",
    "    [24.4481258671796,-18.0324012215159],\n",
    "    [2.82048515404903,-22.4208457598703],\n",
    "    [30.8803004755948,-11.455358009593],\n",
    "    [15.4586738236098,-11.1242825084309],\n",
    "    [28.5332537090494,-14.7898744423126],\n",
    "    [40.4830293441052,-2.41946428697183],\n",
    "    [15.7563759125684,-13.5771266003795],\n",
    "    [19.3635588851727,-20.6224770470434],\n",
    "    [13.4212840786467,-19.0238227375766],\n",
    "    [7.77570680426702,-16.6385739839089],\n",
    "    [21.4865983854408,-15.290799330002],\n",
    "    [12.6392705930724,-23.6433305964301],\n",
    "    [12.4746151388128,-17.9720169566614],\n",
    "    [23.4572410437998,-14.602080545086],\n",
    "    [13.6878189833565,-18.9687408182414],\n",
    "    [15.4077465943441,-14.5352487124086],\n",
    "    [20.3356581548895,-10.0883159703702],\n",
    "    [20.7093833689359,-12.6939091236766],\n",
    "    [11.1032293684441,-14.1383848928755],\n",
    "    [17.5048321498308,-9.2338593361801],\n",
    "    [16.3303688220188,-15.1054735529158],\n",
    "    [26.6929062710726,-13.306030567991],\n",
    "    [34.4985678099711,-9.86199941278607],\n",
    "    [39.1374291499406,-10.5621430853401],\n",
    "    [21.9088956482146,-9.95198845621849],\n",
    "    [22.2367457578087,-17.2200123442707],\n",
    "    [10.0032784145577,-19.3557700653426],\n",
    "    [14.045833906665,-15.871937521131],\n",
    "    [15.5640911917607,-18.3396956121887],\n",
    "    [24.4771926581586,-14.8715313479137],\n",
    "    [26.533415556629,-14.693883922494],\n",
    "    [12.8722580202544,-21.2750596021509],\n",
    "    [24.4768291376862,-15.9592080959207],\n",
    "    [18.2230748567433,-14.6541444069985],\n",
    "    [4.1902148367447,-20.6144032528762],\n",
    "    [12.4332594022086,-16.6079789231489],\n",
    "    [20.5483758651873,-18.8512560786321],\n",
    "    [17.8180560451358,-12.5451990696752],\n",
    "    [11.0071081078049,-20.3938092335862],\n",
    "    [8.30560561422449,-22.9503944138682],\n",
    "    [33.9857852657284,-4.8371294974382],\n",
    "    [17.4376502239652,-14.5095976075022],\n",
    "    [29.0379635148943,-14.8461553663227],\n",
    "    [29.1344666599319,-7.70862921632672],\n",
    "    [32.9730697624544,-15.5839178785654],\n",
    "    [13.4211493998212,-20.150199857584],\n",
    "    [11.380538260355,-12.8619410359766],\n",
    "    [28.672631499186,-8.51866271785711],\n",
    "    [16.4296061111902,-23.3326051279759],\n",
    "    [25.7168371582585,-13.8899296143829],\n",
    "    [13.3185154732595,-17.8959160024249],\n",
    "    [3.60832478605376,-25.4023343597712],\n",
    "    [39.5445949652652,-11.466377647931],\n",
    "    [25.1693484426101,-12.2752652925707],\n",
    "    [25.2884257196471,-7.06710309184533],\n",
    "    [6.77665715793125,-22.3947299635571],\n",
    "    [20.1844223778907,-16.0427471125407],\n",
    "    [25.5506805272535,-9.33856532270204],\n",
    "    [25.1495682602477,-7.17350567090738],\n",
    "    [15.6978431006492,-17.5979197162642],\n",
    "    [37.42780451491,-10.843637288504],\n",
    "    [22.974620174842,-10.6171162611686],\n",
    "    [34.6327117468934,-9.26182440487384],\n",
    "    [34.7042513789061,-6.9630753351114],\n",
    "    [15.6563953929008,-17.2196961218915],\n",
    "    [25.2049825789225,-14.1592086208169]\n",
    "]\n",
    "np_X= np.array(X)\n",
    "plt.plot( np_X[:,0], np_X[:,1],\"ro\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def de_mean_matrix(A):\n",
    "    \"\"\"returns the result of subtracting from every value in A the mean\n",
    "    value of its column. the resulting matrix has mean 0 in every column\"\"\"\n",
    "    nr, nc = shape(A)\n",
    "    column_means, _ = scale(A)\n",
    "    return make_matrix(nr, nc, lambda i, j: A[i][j] - column_means[j])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXwAAAEACAYAAACwB81wAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/j8Qy0ZbFZwqXwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x112938748>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "np_mean_X = np.array(de_mean_matrix(X))\n",
    "\n",
    "plt.plot( np_mean_X[:,0], np_mean_X[:,1],\"ro\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "PCA\n",
      "principal components [[0.92385540904318986, 0.3827416663777809], [-0.38272245395799814, 0.92386336827280258]]\n",
      "first point [0.66637087202545686, 1.6869418499129409]\n",
      "first point transformed [1.2612932692676408, 1.3034686841532084]\n"
     ]
    }
   ],
   "source": [
    "def direction(w):\n",
    "    mag = magnitude(w)\n",
    "    return [w_i / mag for w_i in w]\n",
    "\n",
    "def directional_variance_i(x_i, w):\n",
    "    \"\"\"the variance of the row x_i in the direction w\"\"\"\n",
    "    return dot(x_i, direction(w)) ** 2\n",
    "\n",
    "def directional_variance(X, w):\n",
    "    \"\"\"the variance of the data in the direction w\"\"\"\n",
    "    return sum(directional_variance_i(x_i, w) for x_i in X)\n",
    "\n",
    "def directional_variance_gradient_i(x_i, w):\n",
    "    \"\"\"the contribution of row x_i to the gradient of\n",
    "    the direction-w variance\"\"\"\n",
    "    projection_length = dot(x_i, direction(w))\n",
    "    return [2 * projection_length * x_ij for x_ij in x_i]\n",
    "\n",
    "def directional_variance_gradient(X, w):\n",
    "    return vector_sum(directional_variance_gradient_i(x_i,w) for x_i in X)\n",
    "\n",
    "def first_principal_component(X):\n",
    "    guess = [1 for _ in X[0]]\n",
    "    unscaled_maximizer = maximize_batch(\n",
    "        partial(directional_variance, X),           # is now a function of w\n",
    "        partial(directional_variance_gradient, X),  # is now a function of w\n",
    "        guess)\n",
    "    return direction(unscaled_maximizer)\n",
    "\n",
    "def first_principal_component_sgd(X):\n",
    "    guess = [1 for _ in X[0]]\n",
    "    unscaled_maximizer = maximize_stochastic(\n",
    "        lambda x, _, w: directional_variance_i(x, w),\n",
    "        lambda x, _, w: directional_variance_gradient_i(x, w),\n",
    "        X, [None for _ in X], guess)\n",
    "    return direction(unscaled_maximizer)\n",
    "\n",
    "def project(v, w):\n",
    "    \"\"\"return the projection of v onto w\"\"\"\n",
    "    coefficient = dot(v, w)\n",
    "    return scalar_multiply(coefficient, w)\n",
    "\n",
    "def remove_projection_from_vector(v, w):\n",
    "    \"\"\"projects v onto w and subtracts the result from v\"\"\"\n",
    "    return vector_subtract(v, project(v, w))\n",
    "\n",
    "def remove_projection(X, w):\n",
    "    \"\"\"for each row of X\n",
    "    projects the row onto w, and subtracts the result from the row\"\"\"\n",
    "    return [remove_projection_from_vector(x_i, w) for x_i in X]\n",
    "\n",
    "def principal_component_analysis(X, num_components):\n",
    "    components = []\n",
    "    for _ in range(num_components):\n",
    "        component = first_principal_component(X)\n",
    "        components.append(component)\n",
    "        X = remove_projection(X, component)\n",
    "\n",
    "    return components\n",
    "\n",
    "def transform_vector(v, components):\n",
    "    return [dot(v, w) for w in components]\n",
    "\n",
    "def transform(X, components):\n",
    "    return [transform_vector(x_i, components) for x_i in X]\n",
    "\n",
    "print(\"PCA\")\n",
    "\n",
    "Y = de_mean_matrix(X)\n",
    "components = principal_component_analysis(Y, 2)\n",
    "print(\"principal components\", components)\n",
    "print(\"first point\", Y[0])\n",
    "print(\"first point transformed\", transform_vector(Y[0], components))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3.0
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.4.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}