1-7-Figure 7-Distribution of fraction of daily repeat consumptions.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "from collections import Counter\n",
    "from datetime import datetime\n",
    "\n",
    "import matplotlib.dates as mdates\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "\n",
    "from src.config import FIG_DIR, DATA_DIR\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "figure_data_file = os.path.join(FIG_DIR, 'fig_6_data.csv')\n",
    "\n",
    "if not os.path.exists(figure_data_file):\n",
    "    df = pd.read_csv(os.path.join(DATA_DIR, 'MFP.csv'))\n",
    "    days_ref = np.load(os.path.join(DATA_DIR, 'days_date_mapping.npy')).item()\n",
    "    date_ref = {v: k for k, v in days_ref.items()}\n",
    "    df['date'] = df['days'].apply(lambda s: date_ref[s])\n",
    "\n",
    "    cols = ['uid', 'days', 'food_id']\n",
    "    dic = df[cols].drop_duplicates().groupby(['uid', 'days'])['food_id'].apply(list).to_dict()\n",
    "\n",
    "    def enlist(lst):\n",
    "        \"\"\" Enlist a mixture of list and str, return as original order\"\"\"\n",
    "        res = []\n",
    "        for i in lst:\n",
    "            if isinstance(i, list) or isinstance(i, tuple):\n",
    "                res.extend(i)\n",
    "            else:\n",
    "                if i == i:\n",
    "                    res.append(i)\n",
    "        return res\n",
    "\n",
    "    # times of a food item was consumed by a user in the recent hist\n",
    "    bounded_freq_dic = {}\n",
    "    window_size = 7\n",
    "    for user, t in sorted(dic.keys()):\n",
    "        all_food = enlist([dic.get((user, n), []) for n in np.arange(t - (window_size - 1), t + 1, 1)])\n",
    "        bounded_freq_dic.update({(user, t): dict(Counter(all_food))})\n",
    "\n",
    "    def find_freq(line, ref=bounded_freq_dic):\n",
    "        return ref[(line['uid'], line['days'])][line['food_id']]\n",
    "\n",
    "    df['cnt_bounded'] = df.apply(find_freq, axis=1)\n",
    "    df['is_repeat'] = df['cnt_bounded'] > 1\n",
    "    df['is_repeat'] = df['is_repeat'].astype(int)\n",
    "\n",
    "    # label meal with index\n",
    "    meal_ref = {'breakfast': 0, 'lunch': 1, 'dinner': 2, 'snack': 3}\n",
    "    df['meal_index'] = df['meal'].apply(lambda s: meal_ref[s])\n",
    "\n",
    "    # temporal\n",
    "    data = df.groupby(['date'])['is_repeat'].agg(['sum', 'mean']).reset_index()\n",
    "\n",
    "    data['days'] = data['date'].apply(lambda s: days_ref[s])\n",
    "    data = data.rename(str, columns={'sum': \"n_repeat\", 'mean': \"percentage_repeat_consumption\"})\n",
    "    df_initial_date = '2014-09-14'\n",
    "    date_format = '%Y-%m-%d'\n",
    "    # initial date falls on which weekday\n",
    "    df_offset = datetime.strptime(df_initial_date, date_format).weekday() + 1\n",
    "    # test date after passing 7 training & 1 validation days\n",
    "    data['weekday'] = data['days'].apply(lambda t: (t + df_offset - 1) % 7 + 1)\n",
    "    weekday_ref = {7: 'Sunday', 1: 'Monday', 2: 'Tuesday', 3: 'Wednesday', 4: 'Thursday', 5: 'Friday', 6: 'Saturday'}\n",
    "    data['day'] = data['weekday'].apply(lambda s: weekday_ref[s][:3])\n",
    "    data.to_csv(figure_data_file, index=False)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_fig7(df, col='percentage_repeat_consumption'):\n",
    "    figure_name = os.path.join(FIG_DIR, 'paper_fig_7.png')\n",
    "\n",
    "    plt.rc('pdf', fonttype=42)\n",
    "    sns.set(style=\"ticks\")\n",
    "    sns.set_context('paper', font_scale=1.3)\n",
    "    plt.subplots(figsize=(6, 3))\n",
    "\n",
    "    ax = sns.boxplot(x=\"day\", y=col, data=df, showfliers=False,\n",
    "                     order=[\"Mon\", \"Tue\", \"Wed\", \"Thu\", \"Fri\", \"Sat\", 'Sun'],\n",
    "                     linewidth=2.5, color='darkgrey', )\n",
    "    ax.grid(ls=\"dotted\")\n",
    "    ax.set_ylabel(\"% Repeat consumption\")\n",
    "    ax.set_xlabel(\"Day\")\n",
    "    sns.despine(offset=10, trim=True)\n",
    "    plt.tight_layout()\n",
    "    plt.savefig(figure_name, dpi=200, transparent=False, bbox_inches='tight', pad_inches=0.0)\n",
    "    plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.read_csv(figure_data_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_fig7(data.iloc[7:,])"
   ]
  }
 ],
 "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}