sleep/sleep-research.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Sleep Analysis\n",
    "\n",
    "This is the situation that we will explore.\n",
    "\n",
    "<img src=\"img.png\" width=\"70%\"></img>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pylab as plt\n",
    "df = pd.read_csv(\"/Users/vincent/Downloads/sleep_deprived_coding.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>n</th>\n",
       "      <th>mean_unit_tests</th>\n",
       "      <th>mean_asserts</th>\n",
       "      <th>mean_user_stories</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>sleep</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>deprived</th>\n",
       "      <td>15</td>\n",
       "      <td>1.133333</td>\n",
       "      <td>3.266667</td>\n",
       "      <td>1.600000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>normal</th>\n",
       "      <td>21</td>\n",
       "      <td>2.904762</td>\n",
       "      <td>4.952381</td>\n",
       "      <td>2.714286</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           n  mean_unit_tests  mean_asserts  mean_user_stories\n",
       "sleep                                                         \n",
       "deprived  15         1.133333      3.266667           1.600000\n",
       "normal    21         2.904762      4.952381           2.714286"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(df\n",
    "  .groupby('sleep')\n",
    "  .agg(n=('id', 'count'),\n",
    "       mean_unit_tests=('passed_unit_tests', np.mean),\n",
    "       mean_asserts=('passed_asserts', np.mean),\n",
    "       mean_user_stories=('tackled_user_stories', np.mean),))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def reshuffle(dataf):\n",
    "    return (dataf\n",
    "             .sample(36)\n",
    "             .reset_index(drop=True)\n",
    "             .assign(sleep=lambda d: np.where(d.index < 15, 'deprived', 'normal')))\n",
    "\n",
    "def calc_diff(dataf):\n",
    "    agg = (dataf\n",
    "           .groupby('sleep')\n",
    "           .agg(mean_unit_tests=('passed_unit_tests', np.mean),\n",
    "                mean_asserts=('passed_asserts', np.mean),\n",
    "                mean_user_stories=('tackled_user_stories', np.mean),)).T\n",
    "    return agg['deprived'] - agg['normal']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<img src=\"img2.png\" width=\"90%\"></img>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "n = 1000\n",
    "results = np.zeros((n, 3))\n",
    "for i in range(n):\n",
    "    results[i, :] = calc_diff(reshuffle(df))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_diff = pd.DataFrame(results, \n",
    "                       columns=['diff_unit_tests', \n",
    "                                'diff_asserts', \n",
    "                                'diff_user_stories'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's make a plot of this."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'mean_unit_tests': -1.7714285714285714,\n",
       " 'mean_asserts': -1.685714285714286,\n",
       " 'mean_user_stories': -1.1142857142857143}"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dict(calc_diff(df))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'mean_unit_tests': 0.19480519480519476,\n",
       " 'mean_asserts': 0.29220779220779214,\n",
       " 'mean_user_stories': 0.05844155844155852}"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "agg = (df\n",
    "       .assign(gpa=lambda d: d['gpa'] < d['gpa'].mean())\n",
    "       .groupby('gpa')\n",
    "       .agg(mean_unit_tests=('passed_unit_tests', np.mean),\n",
    "            mean_asserts=('passed_asserts', np.mean),\n",
    "            mean_user_stories=('tackled_user_stories', np.mean))).T\n",
    "\n",
    "effect_dict = dict(agg[False] - agg[True])\n",
    "effect_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12, 4))\n",
    "\n",
    "for idx, name in enumerate(['unit_tests', 'asserts', 'user_stories']):\n",
    "    plt.subplot(131 + idx)\n",
    "    plt.hist(df_diff[f'diff_{name}'])\n",
    "    plt.vlines(effect_dict[f'mean_{name}'], 0, 300, color='red')\n",
    "    plt.title(name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'mean_unit_tests': 0.19480519480519476,\n",
       " 'mean_asserts': 0.29220779220779214,\n",
       " 'mean_user_stories': 0.05844155844155852}"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "effect_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.625"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_diff.assign(limit=lambda d: d['diff_unit_tests'] < 0.19480)['limit'].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.713"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_diff.assign(limit=lambda d: d['diff_asserts'] < 0.29220)['limit'].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.566"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_diff.assign(limit=lambda d: d['diff_user_stories'] < 0.05844)['limit'].mean()"
   ]
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}