top_team_generator.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Team with most points across season to date\n",
    "\n",
    "\n",
    "### Assumptions made\n",
    "- Based on the idea that the players with the CURRENT highest points, will continue to produce the greatest no. of points\n",
    "- Players with <90% of playing have been removed\n",
    "- INCORRECT - Currently doesn't account for squad rules\n",
    "- Assumes budget of £100 million\n",
    "\n",
    "\n",
    "### Selection Procces Method\n",
    "- Initially picks the players with highest points from the following positions\n",
    "-- 2 GKs\n",
    "-- 5 DEFs\n",
    "-- 5 MIDs\n",
    "-- 3 ATTs\n",
    "\n",
    "- Then (assuming the team will value > £100m) the picked player with the lowest delta will be removed\n",
    "- The next_highest player from that position will then fill this place.\n",
    "- These previous two steps will then iterate until the squad value is <= £100m\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import requests\n",
    "import json\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Make a get request to api\n",
    "response = requests.get(\"https://fantasy.premierleague.com/drf/bootstrap-static\")\n",
    "data = response.json()\n",
    "\n",
    "# Print the status code of the response.\n",
    "#print(response.content)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Extract json to pandas dataframe\n",
    "elements = data['elements']\n",
    "ds = pd.DataFrame(elements)\n",
    "\n",
    "#Converts cost into millions\n",
    "ds['now_cost'] = ds['now_cost'] /10\n",
    "\n",
    "#Removes any player with NaN chance of playing\n",
    "ds = ds.dropna(how='any', subset=['chance_of_playing_next_round'])\n",
    "\n",
    "#Removes any player with <90% chance of playing\n",
    "is_likelyToPlay = ds['chance_of_playing_next_round']>90\n",
    "ds = ds[is_likelyToPlay]\n",
    "\n",
    "#Creation of positional datasets\n",
    "\n",
    "is_gk = ds['element_type']==1\n",
    "is_def = ds['element_type']==2\n",
    "is_mid = ds['element_type']==3\n",
    "is_att = ds['element_type']==4\n",
    "\n",
    "gks = ds[is_gk]\n",
    "defs = ds[is_def]\n",
    "mids = ds[is_mid]\n",
    "atts = ds[is_att]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\domin\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:4: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  after removing the cwd from sys.path.\n",
      "C:\\Users\\domin\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:11: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  # This is added back by InteractiveShellApp.init_path()\n",
      "C:\\Users\\domin\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:18: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "C:\\Users\\domin\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:25: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n"
     ]
    }
   ],
   "source": [
    "# Gks add delta\n",
    "z = np.polyfit(gks['now_cost'], gks['total_points'], 1)\n",
    "p = np.poly1d(z)\n",
    "gks['delta'] = gks['total_points'] - p(gks['now_cost'])\n",
    "gks = gks.sort_values(by=['total_points'], ascending=False)\n",
    "gks['pos'] = 'gk'\n",
    "\n",
    "# Def add delta\n",
    "z = np.polyfit(defs['now_cost'], defs['total_points'], 1)\n",
    "p = np.poly1d(z)\n",
    "defs['delta'] = defs['total_points'] - p(defs['now_cost'])\n",
    "defs = defs.sort_values(by=['total_points'], ascending=False)\n",
    "defs['pos'] = 'def'\n",
    "\n",
    "# Mid add delta\n",
    "z = np.polyfit(mids['now_cost'], mids['total_points'], 1)\n",
    "p = np.poly1d(z)\n",
    "mids['delta'] = mids['total_points'] - p(mids['now_cost'])\n",
    "mids = mids.sort_values(by=['total_points'], ascending=False)\n",
    "mids['pos'] = 'mid'\n",
    "\n",
    "# Att Summary\n",
    "z = np.polyfit(atts['now_cost'], atts['total_points'], 1)\n",
    "p = np.poly1d(z)\n",
    "atts['delta'] = atts['total_points'] - p(atts['now_cost'])\n",
    "atts = atts.sort_values(by=['total_points'], ascending=False)\n",
    "atts['pos'] = 'att'\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Dictionary to store the player with the next highest points by position\n",
    "next_highest = {}\n",
    "next_highest['gk'] = 3\n",
    "next_highest['def'] = 6\n",
    "next_highest['mid'] = 6\n",
    "next_highest['att'] = 4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              second_name  total_points      delta  now_cost  pos\n",
      "453                Lloris            80  34.333333       5.4   gk\n",
      "393              Dubravka            79  41.600000       5.0   gk\n",
      "323             Robertson           124  50.228634       6.7  def\n",
      "322              van Dijk           113  45.679808       6.4  def\n",
      "146  Luiz Moreira Marinho           105  48.431765       5.9  def\n",
      "148               Rüdiger            92  33.281374       6.0  def\n",
      "550               Doherty            90  48.484505       5.2  def\n",
      "329                 Salah           162   2.819707      13.4  mid\n",
      "152                Hazard           144  18.733456      11.1  mid\n",
      "350              Sterling           133   4.784435      11.3  mid\n",
      "45                 Fraser           111  59.458997       6.1  mid\n",
      "327                  Mané           108   6.325629       9.5  mid\n",
      "21             Aubameyang           134  24.594427      11.3  att\n",
      "333               Firmino           104  21.030462       9.3  att\n",
      "360                Agüero            99  -9.083771      11.2  att\n",
      "\n",
      "\n",
      "Initial Squad Cost: 123.79999999999998\n",
      "\n",
      "\n",
      "Iteration 1\n",
      "Cut player: Agüero\n",
      "Replacement player: Rashford\n",
      "Updated Squad Cost: 120.10000000000002\n",
      "\n",
      "\n",
      "Iteration 2\n",
      "Cut player: Salah\n",
      "Replacement player: Sigurdsson\n",
      "Updated Squad Cost: 114.00000000000001\n",
      "\n",
      "\n",
      "Iteration 3\n",
      "Cut player: Sterling\n",
      "Replacement player: de Andrade\n",
      "Updated Squad Cost: 109.60000000000002\n",
      "\n",
      "\n",
      "Iteration 4\n",
      "Cut player: Mané\n",
      "Replacement player: Veiga de Carvalho e Silva\n",
      "Updated Squad Cost: 107.7\n",
      "\n",
      "\n",
      "Iteration 5\n",
      "Cut player: Hazard\n",
      "Replacement player: Silva\n",
      "Updated Squad Cost: 105.10000000000001\n",
      "\n",
      "\n",
      "Iteration 6\n",
      "Cut player: Silva\n",
      "Replacement player: Martial\n",
      "Updated Squad Cost: 103.9\n",
      "\n",
      "\n",
      "Iteration 7\n",
      "Cut player: Martial\n",
      "Replacement player: Eriksen\n",
      "Updated Squad Cost: 105.9\n",
      "\n",
      "\n",
      "Iteration 8\n",
      "Cut player: Eriksen\n",
      "Replacement player: Doucouré\n",
      "Updated Squad Cost: 102.50000000000001\n",
      "\n",
      "\n",
      "Iteration 9\n",
      "Cut player: Firmino\n",
      "Replacement player: Vardy\n",
      "Updated Squad Cost: 102.10000000000001\n",
      "\n",
      "\n",
      "Iteration 10\n",
      "Cut player: Vardy\n",
      "Replacement player: Murray\n",
      "Updated Squad Cost: 99.5\n",
      "\n",
      "FINAL SQUAD\n",
      "                   second_name  total_points      delta  now_cost  pos  team\n",
      "21                  Aubameyang           134  24.594427      11.3  att     1\n",
      "386                   Rashford            86  26.822894       7.5  att    14\n",
      "71                      Murray            76  32.684515       6.3  att     3\n",
      "148                    Rüdiger            92  33.281374       6.0  def     6\n",
      "322                   van Dijk           113  45.679808       6.4  def    12\n",
      "146       Luiz Moreira Marinho           105  48.431765       5.9  def     6\n",
      "550                    Doherty            90  48.484505       5.2  def    20\n",
      "323                  Robertson           124  50.228634       6.7  def    12\n",
      "453                     Lloris            80  34.333333       5.4   gk    17\n",
      "393                   Dubravka            79  41.600000       5.0   gk    15\n",
      "356  Veiga de Carvalho e Silva            98  24.341335       7.6  mid    13\n",
      "499                   Doucouré            80  31.408019       5.9  mid    18\n",
      "213                 Sigurdsson           102  32.764867       7.3  mid     8\n",
      "219                 de Andrade            99  35.662911       6.9  mid     8\n",
      "45                      Fraser           111  59.458997       6.1  mid     2\n",
      "\n",
      "\n",
      "Final Squad Cost:\n",
      "99.5\n"
     ]
    }
   ],
   "source": [
    "#Dictionary to store the player with the next highest points by position\n",
    "next_highest = {}\n",
    "next_highest['gk'] = 3\n",
    "next_highest['def'] = 6\n",
    "next_highest['mid'] = 6\n",
    "next_highest['att'] = 4\n",
    "\n",
    "squad = pd.DataFrame()\n",
    "squad = pd.concat([squad, gks.head(2)])\n",
    "squad = pd.concat([squad, defs.head(5)])\n",
    "squad = pd.concat([squad, mids.head(5)])\n",
    "squad = pd.concat([squad, atts.head(3)])\n",
    "\n",
    "squad_cost = sum(squad['now_cost'])\n",
    "print(squad[['second_name','total_points','delta','now_cost', 'pos']])\n",
    "print('\\n')\n",
    "print('Initial Squad Cost: ' + str(squad_cost))\n",
    "\n",
    "iteration = 1\n",
    "\n",
    "while (squad_cost > 100):\n",
    "    print('\\n\\nIteration ' + str(iteration))\n",
    "    #Removes player with lowest delta\n",
    "    cut_player = squad.sort_values(by=['delta'], ascending=True).head(1)\n",
    "    squad = squad.sort_values(by=['delta'], ascending=True).iloc[1:]\n",
    "    print('Cut player: ' + str(cut_player[['second_name']].iloc[0]['second_name']))\n",
    "    \n",
    "    if (cut_player[['pos']].iloc[0]['pos'] == 'gk'):\n",
    "        replacement_player = gks.iloc[[next_highest['gk']]]\n",
    "        next_highest['gk'] += 1\n",
    "        \n",
    "    elif (cut_player[['pos']].iloc[0]['pos'] == 'def'):\n",
    "        replacement_player = defs.iloc[[next_highest['def']]]\n",
    "        next_highest['def'] += 1\n",
    "        \n",
    "    elif (cut_player[['pos']].iloc[0]['pos'] == 'mid'):\n",
    "        replacement_player = mids.iloc[[next_highest['mid']]]\n",
    "        next_highest['mid'] += 1\n",
    "\n",
    "    elif (cut_player[['pos']].iloc[0]['pos'] == 'att'):\n",
    "        replacement_player = atts.iloc[[next_highest['att']]]\n",
    "        next_highest['att'] += 1\n",
    "        \n",
    "        \n",
    "    print('Replacement player: ' + str(replacement_player.iloc[0]['second_name']))\n",
    "    squad = pd.concat([squad, replacement_player.head(1)])\n",
    "    \n",
    "    \n",
    "    squad_cost = sum(squad['now_cost'])\n",
    "    print('Updated Squad Cost: ' + str(squad_cost))\n",
    "    iteration += 1\n",
    "    \n",
    "print('\\nFINAL SQUAD')\n",
    "squad = squad.sort_values(by=['pos'], ascending=True)\n",
    "print(squad[['second_name','total_points','delta','now_cost', 'pos','team']])\n",
    "print('\\n')\n",
    "squad_cost = sum(squad['now_cost'])\n",
    "print('Final Squad Cost:\\n' + str(squad_cost))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}