Filtering and summarizing expert rankings.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Filtering and summarizing expert rankings\n",
    "\n",
    "Now that all of the expert ranking data is downloaded and saved, I use pandas to calculate mean/median rankings and then rank by position. Finally, it's all saved to a spreadsheet."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "import seaborn as sns\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import datetime"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Filtering experts and then trimming down the rankings pool"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "TODAYS_DATE = datetime.date.today().strftime(\"%Y_%m_%d\")\n",
    "SCORING_TYPE = 'ppr'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "         date  draft_rank  expert_id  in_season_rank             name  \\\n",
      "0  2017-08-23       108.0          3           104.0    Eric Karabell   \n",
      "1  2017-08-23         NaN          5             NaN  Staff Composite   \n",
      "2  2017-08-25        69.0          7            65.0     Andy Behrens   \n",
      "3  2017-08-25        28.0          9            66.0  Scott Pianowski   \n",
      "4  2017-08-23        76.0         15             8.0   Staff Rankings   \n",
      "\n",
      "          source  \n",
      "0           ESPN  \n",
      "1           ESPN  \n",
      "2  Yahoo! Sports  \n",
      "3  Yahoo! Sports  \n",
      "4   ScoutFantasy  \n"
     ]
    }
   ],
   "source": [
    "expert_list_file = 'data/fp_experts_{}_{}.tsv'.format(SCORING_TYPE, TODAYS_DATE)\n",
    "expert_df = pd.read_csv(expert_list_file,sep='\\t')\n",
    "print(expert_df.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First we'll compute how \"old\" each ranking is."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "        date  draft_rank  expert_id  in_season_rank             name  \\\n",
      "0 2017-08-23       108.0          3           104.0    Eric Karabell   \n",
      "1 2017-08-23         NaN          5             NaN  Staff Composite   \n",
      "2 2017-08-25        69.0          7            65.0     Andy Behrens   \n",
      "3 2017-08-25        28.0          9            66.0  Scott Pianowski   \n",
      "4 2017-08-23        76.0         15             8.0   Staff Rankings   \n",
      "\n",
      "          source  days_ago  \n",
      "0           ESPN         2  \n",
      "1           ESPN         2  \n",
      "2  Yahoo! Sports         0  \n",
      "3  Yahoo! Sports         0  \n",
      "4   ScoutFantasy         2  \n"
     ]
    }
   ],
   "source": [
    "expert_df['date'] = pd.to_datetime(expert_df['date'])\n",
    "expert_df['days_ago'] = (datetime.date.today() - expert_df['date']).dt.days\n",
    "print(expert_df.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Then we'll filter on rankings age. We could also filter based on each expert's overall ranking from last year, etc., but I think the more data the better."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "82\n"
     ]
    }
   ],
   "source": [
    "filtered_experts = expert_df[expert_df['days_ago'] < 8][['expert_id','days_ago']]\n",
    "print(len(filtered_experts))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   expert_id        player_name position  rank team\n",
      "0          3      David Johnson       RB     1  ARI\n",
      "1          3       Le'Veon Bell       RB     2  PIT\n",
      "2          3      Antonio Brown       WR     3  PIT\n",
      "3          3  Odell Beckham Jr.       WR     4  NYG\n",
      "4          3         Mike Evans       WR     5   TB\n",
      "21793\n"
     ]
    }
   ],
   "source": [
    "expert_rankings_file = 'data/fp_rankings_{}_{}.tsv'.format(SCORING_TYPE,TODAYS_DATE)\n",
    "all_rankings = pd.read_csv(expert_rankings_file, sep='\\t')\n",
    "print(all_rankings.head())\n",
    "print(len(all_rankings))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that we have all of the rankings read in, we'll filter out the old experts."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "20358\n"
     ]
    }
   ],
   "source": [
    "filtered_expert_ids = filtered_experts['expert_id'].tolist()\n",
    "filtered_rankings = all_rankings[all_rankings['expert_id'].isin(filtered_expert_ids)]\n",
    "print(len(filtered_rankings))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next we'll filter again based on how many rankings each player has across all the expert lists."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "player_name      team  position\n",
      "A.J. Derby       DEN   TE          14\n",
      "A.J. Green       CIN   WR          82\n",
      "Aaron Jones      GB    RB          20\n",
      "Aaron Ripkowski  GB    RB           1\n",
      "Aaron Rodgers    GB    QB          82\n",
      "dtype: int64\n",
      "459\n"
     ]
    }
   ],
   "source": [
    "ranking_counts = filtered_rankings.groupby(['player_name','team','position']).size()\n",
    "print(ranking_counts.head())\n",
    "print(len(ranking_counts))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are 459 different players that have at least 1 ranking. Let's make a KDE plot to see what a cut-off value is."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAFVCAYAAAD2VHb/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x110e39ba8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.kdeplot(ranking_counts, shade=True);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Based on this plot, it looks like appearing in 40 expert rankings is a reasonable cut-off."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18510\n"
     ]
    }
   ],
   "source": [
    "RANKING_CUTOFF = 40\n",
    "filtered_rankings = filtered_rankings.groupby(['player_name','team','position']).filter(lambda x: len(x) > RANKING_CUTOFF)\n",
    "print(len(filtered_rankings))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Ok, we have a pretty good dataset now, and we're only down to 18,510 observations (from 21,793 originally)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "filtered_rankings_file = 'data/fp_rankings_{}_{}_filtered.tsv'.format(SCORING_TYPE,TODAYS_DATE)\n",
    "filtered_rankings.to_csv(filtered_rankings_file, sep='\\t', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Computing summary statistics on rankings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "           player_name position team  median_rank  mean_rank\n",
      "64       David Johnson       RB  ARI          1.0   1.195122\n",
      "158       Le'Veon Bell       RB  PIT          2.0   1.841463\n",
      "14       Antonio Brown       WR  PIT          3.0   3.475610\n",
      "141        Julio Jones       WR  ATL          5.0   5.682927\n",
      "194  Odell Beckham Jr.       WR  NYG          5.0   5.756098\n"
     ]
    }
   ],
   "source": [
    "overall_ranks = filtered_rankings.groupby(['player_name', 'position','team'],\n",
    "                                     as_index=False).agg({'rank':[np.median, np.mean]})\n",
    "overall_ranks.columns = ['player_name','position','team','median_rank','mean_rank']\n",
    "overall_ranks.sort_values(['median_rank','mean_rank'], ascending=True, inplace=True)\n",
    "print(overall_ranks.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "     position_rank\n",
      "64               1\n",
      "158              2\n",
      "14               1\n",
      "141              2\n",
      "194              3\n"
     ]
    }
   ],
   "source": [
    "position_ranks = overall_ranks.groupby('position')['median_rank'].rank(method='first')\n",
    "position_ranks = position_ranks.to_frame(name='position_rank')\n",
    "position_ranks['position_rank'] = position_ranks['position_rank'].astype(int)\n",
    "print(position_ranks.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "           player_name position team  median_rank  mean_rank  position_rank\n",
      "64       David Johnson       RB  ARI          1.0   1.195122              1\n",
      "158       Le'Veon Bell       RB  PIT          2.0   1.841463              2\n",
      "14       Antonio Brown       WR  PIT          3.0   3.475610              1\n",
      "141        Julio Jones       WR  ATL          5.0   5.682927              2\n",
      "194  Odell Beckham Jr.       WR  NYG          5.0   5.756098              3\n",
      "160       LeSean McCoy       RB  BUF          6.0   6.073171              3\n",
      "179      Melvin Gordon       RB  LAC          8.0   8.341463              4\n",
      "182         Mike Evans       WR   TB          9.0   9.182927              4\n",
      "0           A.J. Green       WR  CIN          9.0   9.475610              5\n",
      "78     Devonta Freeman       RB  ATL          9.0   9.926829              5\n"
     ]
    }
   ],
   "source": [
    "overall_ranks = overall_ranks.join(position_ranks)\n",
    "print(overall_ranks.head(10))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Saving to a spreadsheet"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, we'll save it to a machine-readable TSV file before moving on to a human-readable spreadsheet."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "overall_rankings_file = 'data/fp_rankings_{}_{}_summary.tsv'.format(SCORING_TYPE,TODAYS_DATE)\n",
    "overall_ranks.to_csv(overall_rankings_file, sep='\\t', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "filename = 'rankings_summary_{}_{}.xlsx'.format(SCORING_TYPE,TODAYS_DATE)\n",
    "writer = pd.ExcelWriter(filename, engine='xlsxwriter')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "overall_ranks = overall_ranks[['player_name','position','position_rank','team','median_rank','mean_rank']]\n",
    "overall_ranks.to_excel(writer,sheet_name='Overall', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "for position in ['WR','RB','TE','QB','D/ST','K']:\n",
    "    position_ranks = overall_ranks[overall_ranks['position'] == position]\n",
    "    if position == 'D/ST': # Since player_name and team are redundant\n",
    "        position_ranks = position_ranks[['player_name','median_rank','mean_rank']]\n",
    "        # Excel doesn't allow slashes in sheetnames\n",
    "        position_ranks.to_excel(writer, sheet_name='DEF', index=False)\n",
    "    else:\n",
    "        position_ranks = position_ranks[['player_name','team','median_rank','mean_rank']]\n",
    "        position_ranks.to_excel(writer, sheet_name=position, index=False)\n",
    "writer.save()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [default]",
   "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}