halite_stats.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Requirements\n",
    "- matplotlib\n",
    "- numpy\n",
    "- pandas\n",
    "- seaborn\n",
    "- scipy\n",
    "- requests\n",
    "- (optional) tqdm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pathlib\n",
    "import sys\n",
    "PATH = pathlib.Path().absolute().resolve()\n",
    "\n",
    "%matplotlib inline\n",
    "import math\n",
    "import json\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.ticker as tk\n",
    "import matplotlib.colors as clr\n",
    "import seaborn as sns\n",
    "    \n",
    "sns.set(style=\"whitegrid\", font_scale=1.4)\n",
    "COLORS = sns.color_palette()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from scipy.signal import savgol_filter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "import requests\n",
    "try:\n",
    "    from tqdm import tqdm\n",
    "except ImportError:\n",
    "    # tqdm placeholder\n",
    "    class tqdm:\n",
    "        def __init__(self, it=None): \n",
    "            self.it=it\n",
    "        def __enter__(self): \n",
    "            return self\n",
    "        def __exit__(self, *args, **kwargs): \n",
    "            pass\n",
    "        def __iter__(self): \n",
    "            return iter(self.it)\n",
    "        def update(*args): \n",
    "            pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "leaderbord_url = lambda offset, limit=100: f\"https://api.2018.halite.io/v1/api/leaderboard?offset={offset}&limit={limit}\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def ax_style(ax):\n",
    "    ax.grid(which='minor', axis='x', dashes=(5, 5), linewidth=.5)\n",
    "    ax.grid(which='minor', axis='y', dashes=(5, 5), linewidth=.5)\n",
    "    ax.legend()\n",
    "    ax.minorticks_on()\n",
    "    return ax"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "leaderboard_data = []\n",
    "chunk = True\n",
    "step = 100\n",
    "i = 0\n",
    "with tqdm() as progress:\n",
    "    while chunk:\n",
    "        url = leaderbord_url(step * i, step)\n",
    "        i += 1\n",
    "        chunk = requests.get(url).json()\n",
    "        leaderboard_data.extend(chunk)\n",
    "        progress.update(step)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "MU_CUT = 30\n",
    "SIGMA_CUT = 2\n",
    "df = pd.DataFrame(leaderboard_data)\n",
    "df = df[df.mu >= MU_CUT]\n",
    "df_filtered = df[df.sigma <= SIGMA_CUT]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "fig = plt.figure(figsize=(20, 10))\n",
    "ax = fig.add_subplot(111)\n",
    "df_filtered.plot(x=\"rank\", y=\"mu\", ax=ax, marker=',', linestyle=' ', label=r\"$\\mu$ raw\")\n",
    "Y_smooth = savgol_filter(df.mu, 101, 3)\n",
    "ax.plot(df[\"rank\"], Y_smooth, label=r\"$\\mu$\")\n",
    "\n",
    "ax.set_xlabel(\"Ladder position\")\n",
    "ax.set_ylabel(r\"$\\mu$\")\n",
    "ax.set_xlim(len(df), 0)\n",
    "ax_style(ax)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "fig = plt.figure(figsize=(20, 10))\n",
    "ax1 = fig.add_subplot(121)\n",
    "ax2 = fig.add_subplot(122)\n",
    "for i, (df_, ax) in enumerate(((df, ax1), (df[df[\"rank\"] <= 100], ax2))):\n",
    "    values = df_.language.value_counts()[::-1]\n",
    "    values.plot(ax=ax, kind='barh', logx=not i)\n",
    "    ax_style(ax)\n",
    "    ax.set_xlim(not i, values.max() * 1.1)\n",
    "ax1.set_title(\"All players\")\n",
    "ax2.set_title(\"Top100 players\")\n",
    "fig.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "fig = plt.figure(figsize=(20, 30))\n",
    "ax1 = plt.subplot2grid((3, 1), (0, 0))\n",
    "ax2 = plt.subplot2grid((3, 1), (1, 0))\n",
    "rank_cut = 200\n",
    "top_100 = {}\n",
    "for lang in df.language.unique():\n",
    "    df_ = df[df.language == lang]\n",
    "    if len(df_) >= 30:\n",
    "        sns.distplot(df_.score, ax=ax1, hist=False, label=lang)\n",
    "        sns.distplot(df_[\"rank\"], ax=ax2, hist=False, bins=20, label=lang)\n",
    "        \n",
    "    values = df_[df_[\"rank\"] < rank_cut][\"rank\"]\n",
    "    if len(values) >= 3:\n",
    "        top_100[lang] = values\n",
    "        \n",
    "sns.distplot(df[df.is_gpu_enabled == True][\"rank\"], ax=ax2, kde_kws={\"ls\": \"--\"}, hist=False, label=\"with GPU\")\n",
    "        \n",
    "top_100 = sorted(top_100.items(), reverse=True,\n",
    "                 key=lambda k_v: df[(df.language == k_v[0]) & (df[\"rank\"] <= rank_cut)].score.mean())\n",
    "top_100_k, top_100_v = zip(*top_100)\n",
    "\n",
    "ax_style(ax1)\n",
    "ax1.set_xlim(df.score.max(), MU_CUT)\n",
    "ax1.set_xlabel(r\"Score = $\\mu - 3\\sigma$\")\n",
    "ax1.set_ylabel(\"Distribution\")\n",
    "ax1.set_title(\"Distrubution of score by prorg. lang.\")\n",
    "\n",
    "ax_style(ax2)\n",
    "ax2.set_xlim(1, df[\"rank\"].max())\n",
    "ax2.set_xlabel(r\"Rank\")\n",
    "ax2.set_ylabel(\"Distribution\")\n",
    "ax2.set_title(\"Distrubution of rank by prorg. lang.\")\n",
    "\n",
    "with sns.axes_style(style=\"darkgrid\"):\n",
    "    ax3 = plt.subplot2grid((3, 1), (2, 0))\n",
    "    ax3.hist(\n",
    "        top_100_v, \n",
    "        bins=np.arange(0, rank_cut + rank_cut // 10, 10), \n",
    "        stacked=True, label=top_100_k,\n",
    "        color=sns.color_palette(\"Spectral\", len(top_100_k)),\n",
    "        zorder=-1\n",
    "    )\n",
    "    \n",
    "    ax_style(ax3)\n",
    "    ax3.set_xlim(1, rank_cut + 40)\n",
    "    ax3.set_xlabel(f\"Rank (top{rank_cut})\")\n",
    "    ax3.set_ylabel(\"Distribution\")\n",
    "    ax3.set_title(f\"Distrubution of rank (top{rank_cut}) by prorg. lang.\")\n",
    "        \n",
    "\n",
    "fig.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}

requirements.txt

matplotlib
numpy
pandas
seaborn
scipy
requests
tqdm