knn.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 437,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd, numpy as np, matplotlib, matplotlib.pyplot as plt\n",
    "from sklearn import preprocessing\n",
    "import numpy as np\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from scipy.spatial import distance\n",
    "import math\n",
    "\n",
    "initYear = '2016'\n",
    "years = [initYear, str(int(initYear) + 1), str(int(initYear) + 2)]\n",
    "lastYear = int(years[2])\n",
    "\n",
    "\n",
    "# use this read csv for 2017-2019 predictions\n",
    "# allData = pd.read_csv(str(lastYear + 1) + 'PredictionsAndActual.csv').dropna()\n",
    "\n",
    "# use this for kenny's stuff\n",
    "allData = pd.read_csv('allPredictions2019.csv').dropna()\n",
    "# allData = pd.read_csv(str(lastYear + 1) + 'PredictionsAndActual.csv')\n",
    "# data = pd.read_csv('allPredictions2019.csv')\n",
    "# data = data[['playerid', 'HuberRegressorRBI', 'HuberRegressorAVG', 'HuberRegressorHR', 'HuberRegressorOPS', 'HuberRegressorRuns', 'RidgeRBI', 'RidgeAVG', 'RidgeHR', 'RidgeOPS', 'RidgeRuns']]\n",
    "# data = data.join(allData, how='outer', lsuffix='playerid', rsuffix='playerid')\n",
    "# # data = data.rename(columns={'playeridplayerid': 'playerid'})#, 'HRsPredicted': 'HR', 'RBIsPredicted': 'RBI', 'AVGPredicted': 'AVG', 'opsPredicted': 'OPS' })\n",
    "# allData = data\n",
    "# allData = allData.dropna()\n",
    "# print(allData['playeridplayerid'])\n",
    "\n",
    "\n",
    "# print(list(allData.columns))\n",
    "# print(allData['playerid'])\n",
    "# use this read csv if you want to see what will happen in 2020\n",
    "# allData = pd.read_csv('2020Predictions.csv').dropna()\n",
    "\n",
    "# use this read csv is you want to see how our old predictions fare\n",
    "# allData = pd.read_csv('oldPredictions2019.csv').dropna()\n",
    "\n",
    "def computeData(toScale, colList, dataMode):\n",
    "    if dataMode == 'scale':\n",
    "        return preprocessing.scale(toScale[colList])\n",
    "    else:\n",
    "        return toScale[colList]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 438,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def getDataByLAP(fileName, mode, dataMode):\n",
    "    positionData = pd.read_csv(fileName)\n",
    "    position = positionData.loc[positionData['Season'] == lastYear]\n",
    "    position = position[['playerid', 'Team']].copy()\n",
    "    position = pd.merge(position, allData, on='playerid').dropna()\n",
    "    \n",
    "    if mode == 'predicted':\n",
    "        colList = ['runsPredicted', 'HRsPredicted', 'RBIsPredicted', 'AVGPredicted', 'opsPredicted']\n",
    "    elif mode == 'linearSVR':\n",
    "        colList = ['linearSVRRuns', 'linearSvrHR', 'linearSvrRBI', 'linearSvrAVG', 'linearSvrOPS']\n",
    "    elif mode == 'svr':\n",
    "        colList = ['svrRuns', 'SvrHR', 'svrRBI', 'svrAVG', 'svrOPS']\n",
    "    elif mode == 'huber':\n",
    "        colList = ['HuberRegressorRBI', 'HuberRegressorAVG', 'HuberRegressorHR', 'HuberRegressorOPS', 'HuberRegressorRuns']\n",
    "    elif mode == 'ridge':\n",
    "        colList = ['RidgeRBI', 'RidgeAVG', 'RidgeHR', 'RidgeOPS', 'RidgeRuns']\n",
    "    else:\n",
    "        if lastYear == 2018 or lastYear == 2019:\n",
    "            colList = ['runsActual', 'HRsActual', 'RBIsActuals', 'AVGActuals', 'OpsActuals']\n",
    "        else:\n",
    "            colList = ['runs', 'hr', 'rbi', 'avg', 'ops']\n",
    "            \n",
    "    position[colList] = computeData(position, colList, dataMode)\n",
    "\n",
    "    return position"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 439,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def getMin(wins):\n",
    "    minVal = 100\n",
    "    minRow = []\n",
    "    for index, val in enumerate(wins):\n",
    "        if val['Distance'] < minVal:\n",
    "            minRow = val\n",
    "    print(minRow['Name'])\n",
    "    return minRow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 440,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def compute_by_position(dataMode):\n",
    "    for league in winners:\n",
    "        for position in winners[league]:\n",
    "            if position == 'of':\n",
    "                continue\n",
    "            fileName = winners[league][position]['file']\n",
    "\n",
    "            for year in years: #winners[league][position]['stats']:\n",
    "                data = pd.read_csv(fileName)\n",
    "                data = data.loc[data['Season'] == int(year)]\n",
    "                colList = ['R', 'HR', 'RBI', 'AVG', 'OPS']\n",
    "                data[colList] = computeData(data, colList, dataMode)\n",
    "\n",
    "                if position != 'ofs':\n",
    "                    data['Award'] = np.where(data['Name'] == winners[league][position]['stats'][year]['name'], 1, 0)\n",
    "                    winners[league][position]['stats'][year]['playerdata'] = data[['Season', 'Name', 'playerid', 'R', 'HR', 'RBI', 'AVG', 'OPS', 'Award']]\n",
    "                    data = data.loc[data['Name'] == winners[league][position]['stats'][year]['name']]\n",
    "                    winner = data.loc[data['Name'] == winners[league][position]['stats'][year]['name']]\n",
    "                    winner = winner[['R', 'HR', 'RBI', 'AVG', 'OPS']].copy()\n",
    "                    winners[league][position]['stats'][year]['data'] = winner\n",
    "                else:\n",
    "                    for index, row in data.iterrows():\n",
    "                        if row['Name'] in winners[league][position]['stats'][year]['names']:\n",
    "                            data.at[index, 'Award'] = 1\n",
    "                        else:\n",
    "                            data.at[index, 'Award'] = 0\n",
    "                    winners[league][position]['stats'][year]['playerdata'] = data[['Season', 'Name', 'playerid', 'R', 'HR', 'RBI', 'AVG', 'OPS', 'Award']]\n",
    "                    winner = data.loc[data['Award'] == 1]\n",
    "                    winner = winner[['R', 'HR', 'RBI', 'AVG', 'OPS']].copy()\n",
    "                    winners[league][position]['stats'][year]['data'] = winner\n",
    "            winners[league][position]['average'] = pd.concat((winners[league][position]['stats'][years[0]]['data'], winners[league][position]['stats'][years[1]]['data']))\n",
    "            winners[league][position]['average'] = pd.concat((winners[league][position]['average'], winners[league][position]['stats'][years[2]]['data']))\n",
    "            winners[league][position]['average'] = winners[league][position]['average'].mean(axis=0)\n",
    "            winners[league][position]['average'] = pd.DataFrame(data=winners[league][position]['average']).T\n",
    "            frames = [winners[league][position]['stats'][years[0]]['playerdata'], winners[league][position]['stats'][years[1]]['playerdata'], winners[league][position]['stats'][years[2]]['playerdata']]\n",
    "            winners[league][position]['playerdata'] = pd.concat(frames)\n",
    "        winners[league]['of'] = {}\n",
    "        of1 = pd.DataFrame(data=winners[league]['of1']['average'])\n",
    "        of2 = pd.DataFrame(data=winners[league]['of2']['average'])\n",
    "        of3 = pd.DataFrame(data=winners[league]['of3']['average'])\n",
    "        winners[league]['of']['average'] = pd.concat((of1, of2))\n",
    "        winners[league]['of']['average'] = pd.concat((winners[league]['of']['average'], of3))\n",
    "        winners[league]['of']['average'] = winners[league]['of']['average'].mean(axis=0)\n",
    "        winners[league]['of']['average'] = pd.DataFrame(data=winners[league]['of']['average']).T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 441,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def knn(mode, neighbors, debug):\n",
    "    for league in winners:\n",
    "        otherLeague = 'al'\n",
    "        if league == 'al':\n",
    "            otherLeague = 'nl'\n",
    "        predWinners[league] = {}\n",
    "        for position in winners[league]:\n",
    "            frames = []\n",
    "            temp = []\n",
    "            if position == 'of' or position == 'of1' or position == 'of2' or position == 'of3':\n",
    "                continue\n",
    "\n",
    "            if position != 'dh':\n",
    "                if position == 'ofs':\n",
    "                    for year in years:\n",
    "                        frames.append(pd.DataFrame(data=winners[otherLeague]['of1']['stats'][year]['data']))\n",
    "                        frames.append(pd.DataFrame(data=winners[otherLeague]['of2']['stats'][year]['data']))\n",
    "                        frames.append(pd.DataFrame(data=winners[otherLeague]['of3']['stats'][year]['data']))\n",
    "                    temp = pd.concat(frames)\n",
    "                    temp['Award'] = 1\n",
    "                    train_x = pd.concat((winners[league][position]['playerdata'][['R', 'HR', 'RBI', 'AVG', 'OPS']],temp[['R', 'HR', 'RBI', 'AVG', 'OPS']]), axis=0)\n",
    "                    train_y = pd.concat((winners[league][position]['playerdata'][['Award']],temp[['Award']]))\n",
    "                else:\n",
    "                    for year in years: #winners[otherLeague][position]['stats']:\n",
    "                        frames.append(pd.DataFrame(data=winners[otherLeague][position]['stats'][year]['data']))\n",
    "                    temp = pd.concat(frames)\n",
    "                    temp['Award'] = 1\n",
    "                    train_x = pd.concat((winners[league][position]['playerdata'][['R', 'HR', 'RBI', 'AVG', 'OPS']],temp[['R', 'HR', 'RBI', 'AVG', 'OPS']]), axis=0)\n",
    "                    train_y = pd.concat((winners[league][position]['playerdata'][['Award']],temp[['Award']]))\n",
    "            else:\n",
    "                train_x = winners[league][position]['playerdata'][['R', 'HR', 'RBI', 'AVG', 'OPS']]\n",
    "                train_y = winners[league][position]['playerdata'][['Award']]\n",
    "                \n",
    "            for i, p in enumerate(specificData[league]):\n",
    "                numWinners = 0\n",
    "                awardWinners = []\n",
    "                \n",
    "                if specificData[league][i]['position'] == position:\n",
    "                    if debug:\n",
    "                        print(position)\n",
    "                    predWinners[league][position] = []\n",
    "                    if mode == 'predicted':\n",
    "                        test = specificData[league][i]['data'][['playerid', 'Name', 'runsPredicted', 'HRsPredicted', 'RBIsPredicted', 'AVGPredicted', 'opsPredicted']]\n",
    "                        test = test.rename(columns={'runsPredicted': 'R', 'HRsPredicted': 'HR', 'RBIsPredicted': 'RBI', 'AVGPredicted': 'AVG', 'opsPredicted': 'OPS' })\n",
    "                    elif mode == 'linearSVR':\n",
    "                        test = specificData[league][i]['data'][['playerid', 'Name', 'linearSVRRuns', 'linearSvrHR', 'linearSvrRBI', 'linearSvrAVG', 'linearSvrOPS']]\n",
    "                        test = test.rename(columns={'linearSVRRuns': 'R', 'linearSvrHR': 'HR', 'linearSvrRBI': 'RBI', 'linearSvrAVG': 'AVG', 'linearSvrOPS': 'OPS'})\n",
    "                    elif mode == 'svr':\n",
    "                        test = specificData[league][i]['data'][['playerid', 'Name', 'svrRuns', 'SvrHR', 'svrRBI', 'svrAVG', 'svrOPS']]\n",
    "                        test = test.rename(columns={'svrRuns': 'R', 'SvrHR': 'HR', 'svrRBI': 'RBI', 'svrAVG': 'AVG', 'svrOPS': 'OPS'})\n",
    "                    elif mode == 'huber':\n",
    "                        test = specificData[league][i]['data'][['playerid', 'Name', 'HuberRegressorRBI', 'HuberRegressorAVG', 'HuberRegressorHR', 'HuberRegressorOPS', 'HuberRegressorRuns']]\n",
    "                        test = test.rename(columns={'HuberRegressorRBI': 'RBI', 'HuberRegressorAVG': 'AVG', 'HuberRegressorHR': 'HR', 'HuberRegressorOPS': 'OPS', 'HuberRegressorRuns': 'R'})\n",
    "                    elif mode == 'ridge':\n",
    "                        test = specificData[league][i]['data'][['playerid', 'Name', 'RidgeRBI', 'RidgeAVG', 'RidgeHR', 'RidgeOPS', 'RidgeRuns']]\n",
    "                        test = test.rename(columns={'RidgeRBI': 'RBI', 'RidgeAVG': 'AVG', 'RidgeHR': 'HR', 'RidgeOPS': 'OPS', 'RidgeRuns': 'R'})\n",
    "                    else:\n",
    "                        if lastYear == 2018:\n",
    "                            test = specificData[league][i]['data'][['playerid', 'Name', 'runsActual', 'HRsActual', 'RBIsActuals', 'AVGActuals', 'OpsActuals']]\n",
    "                            test = test.rename(columns={'runsActual': 'R', 'HRsActual': 'HR', 'RBIsActuals': 'RBI', 'AVGActuals': 'AVG', 'OpsActuals': 'OPS'})\n",
    "                        else:\n",
    "                            test = specificData[league][i]['data'][['playerid', 'Name', 'runs', 'hr', 'rbi', 'avg', 'ops']]\n",
    "                            test = test.rename(columns={'runs': 'R', 'hr': 'HR', 'rbi': 'RBI', 'avg': 'AVG', 'ops': 'OPS'})\n",
    "                    classifier = KNeighborsClassifier(n_neighbors=neighbors)\n",
    "                    classifier.fit(train_x, train_y.values.ravel())\n",
    "\n",
    "                    if position != 'dh':\n",
    "                        combinedWinners = pd.concat([winners[league][position]['average'], winners[otherLeague][position]['average']])\n",
    "                        combinedWinners = combinedWinners.mean(axis=0)\n",
    "                    else:\n",
    "                        combinedWinners = winners[league][position]['average']\n",
    "                        \n",
    "                    for index, row in test.iterrows():\n",
    "                        vals = [row['R'], row['HR'], row['RBI'], row['AVG'], row['OPS']]\n",
    "                        test.at[index, 'Distance'] = distance.euclidean([vals], combinedWinners)\n",
    "                    \n",
    "                    test = test.sort_values(by='Distance')\n",
    "                    for index, row in test.iterrows():\n",
    "                        vals = [row['R'], row['HR'], row['RBI'], row['AVG'], row['OPS']]\n",
    "                        res = classifier.predict([vals]) == [1]                        \n",
    "                        if res == 1:\n",
    "                            numWinners = numWinners + 1\n",
    "                            awardWinners.append(row)\n",
    "                            \n",
    "                    if numWinners == 5:\n",
    "                        wins = pd.concat(awardWinners)\n",
    "                        if debug:\n",
    "                            print(wins['Name'])\n",
    "                        predWinners[league][position].append(wins['Name'])\n",
    "                    elif numWinners < 5:\n",
    "                        runs = 5 - numWinners\n",
    "                        if numWinners > 0:\n",
    "                            for row in awardWinners:\n",
    "                                if debug:\n",
    "                                    print(row['Name'])\n",
    "                                predWinners[league][position].append(row['Name'])\n",
    "                                test = test[test['playerid'] != row['playerid']]\n",
    "                        for index in range(runs):\n",
    "                            minRow = test[test['Distance']==test['Distance'].min()]\n",
    "                            if debug:\n",
    "                                print(minRow['Name'].tolist()[0])\n",
    "                            predWinners[league][position].append(minRow['Name'].tolist()[0])\n",
    "                            test = test[test['playerid'] != minRow['playerid'].tolist()[0]]\n",
    "\n",
    "                    else:\n",
    "                        for row in awardWinners:\n",
    "                            if debug:\n",
    "                                print(row['Name'])\n",
    "                            predWinners[league][position].append(row['Name'])\n",
    "                        \n",
    "                if specificData[league][i]['position'] == 'of' and position == 'ofs':\n",
    "                    if debug:\n",
    "                        print(position)\n",
    "                    predWinners[league][position] = []\n",
    "\n",
    "                    if mode == 'predicted':\n",
    "                        test = specificData[league][i]['data'][['playerid', 'Name', 'runsPredicted', 'HRsPredicted', 'RBIsPredicted', 'AVGPredicted', 'opsPredicted']]\n",
    "                        test = test.rename(columns={'runsPredicted': 'R', 'HRsPredicted': 'HR', 'RBIsPredicted': 'RBI', 'AVGPredicted': 'AVG', 'opsPredicted': 'OPS' })\n",
    "                    elif mode == 'linearSVR':\n",
    "                        test = specificData[league][i]['data'][['playerid', 'Name', 'linearSVRRuns', 'linearSvrHR', 'linearSvrRBI', 'linearSvrAVG', 'linearSvrOPS']]\n",
    "                        test = test.rename(columns={'linearSVRRuns': 'R', 'linearSvrHR': 'HR', 'linearSvrRBI': 'RBI', 'linearSvrAVG': 'AVG', 'linearSvrOPS': 'OPS'})\n",
    "                    elif mode == 'svr':\n",
    "                        test = specificData[league][i]['data'][['playerid', 'Name', 'svrRuns', 'SvrHR', 'svrRBI', 'svrAVG', 'svrOPS']]\n",
    "                        test = test.rename(columns={'svrRuns': 'R', 'SvrHR': 'HR', 'svrRBI': 'RBI', 'svrAVG': 'AVG', 'svrOPS': 'OPS'})\n",
    "                    elif mode == 'huber':\n",
    "                        test = specificData[league][i]['data'][['playerid', 'Name', 'HuberRegressorRBI', 'HuberRegressorAVG', 'HuberRegressorHR', 'HuberRegressorOPS', 'HuberRegressorRuns']]\n",
    "                        test = test.rename(columns={'HuberRegressorRBI': 'RBI', 'HuberRegressorAVG': 'AVG', 'HuberRegressorHR': 'HR', 'HuberRegressorOPS': 'OPS', 'HuberRegressorRuns': 'R'})\n",
    "                    elif mode == 'ridge':\n",
    "                        test = specificData[league][i]['data'][['playerid', 'Name', 'RidgeRBI', 'RidgeAVG', 'RidgeHR', 'RidgeOPS', 'RidgeRuns']]\n",
    "                        test = test.rename(columns={'RidgeRBI': 'RBI', 'RidgeAVG': 'AVG', 'RidgeHR': 'HR', 'RidgeOPS': 'OPS', 'RidgeRuns': 'R'})\n",
    "\n",
    "                    else:\n",
    "                        if lastYear == 2018:\n",
    "                            test = specificData[league][i]['data'][['playerid', 'Name', 'runsActual', 'HRsActual', 'RBIsActuals', 'AVGActuals', 'OpsActuals']]\n",
    "                            test = test.rename(columns={'runsActual': 'R', 'HRsActual': 'HR', 'RBIsActuals': 'RBI', 'AVGActuals': 'AVG', 'OpsActuals': 'OPS'})\n",
    "                        else:\n",
    "                            test = specificData[league][i]['data'][['playerid', 'Name', 'runs', 'hr', 'rbi', 'avg', 'ops']]\n",
    "                            test = test.rename(columns={'runs': 'R', 'hr': 'HR', 'rbi': 'RBI', 'avg': 'AVG', 'ops': 'OPS'})\n",
    "\n",
    "                    classifier = KNeighborsClassifier(n_neighbors=neighbors)\n",
    "                    classifier.fit(train_x, train_y.values.ravel())\n",
    "\n",
    "                    if position != 'dh':\n",
    "                        combinedWinners = pd.concat([winners[league][position]['average'], winners[otherLeague][position]['average']])\n",
    "                        combinedWinners = combinedWinners.mean(axis=0)\n",
    "                    else:\n",
    "                        combinedWinners = winners[league][position]['average']\n",
    "                        \n",
    "                    for index, row in test.iterrows():\n",
    "                        vals = [row['R'], row['HR'], row['RBI'], row['AVG'], row['OPS']]\n",
    "                        test.at[index, 'Distance'] = distance.euclidean([vals], combinedWinners)\n",
    "\n",
    "                    for index, row in test.iterrows():\n",
    "                        vals = [row['R'], row['HR'], row['RBI'], row['AVG'], row['OPS']]\n",
    "                        res = classifier.predict([vals]) == [1]\n",
    "                        if res == 1:\n",
    "                            numWinners = numWinners + 1\n",
    "                            awardWinners.append(row)\n",
    "\n",
    "                    if numWinners == 15:\n",
    "                        wins = pd.concat(awardWinners)\n",
    "                        if debug:\n",
    "                            print(wins['Name'])\n",
    "                        predWinners[league][position].append(wins['Name'])\n",
    "                    elif numWinners > 15:\n",
    "                        for row in awardWinners:\n",
    "                            if debug:\n",
    "                                print(row['Name'])\n",
    "                            predWinners[league][position].append(wins['Name'])\n",
    "                    else:\n",
    "                        runs = 15 - numWinners\n",
    "                        if numWinners > 0:\n",
    "                            for row in awardWinners:\n",
    "                                if debug:\n",
    "                                    print(row['Name'])\n",
    "                                predWinners[league][position].append(row['Name'])\n",
    "                                test = test[test['playerid'] != row['playerid']]\n",
    "                        for index in range(runs):\n",
    "                            minRow = test[test['Distance']==test['Distance'].min()]\n",
    "                            if debug:\n",
    "                                print(minRow['Name'].tolist()[0])\n",
    "                            predWinners[league][position].append(minRow['Name'].tolist()[0])\n",
    "                            test = test[test['playerid'] != minRow['playerid'].tolist()[0]]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 442,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "specificData = {\n",
    "    'al': [\n",
    "        {'position': '1b', 'fileName': 'al-1b+2014.csv', 'data': []},\n",
    "        {'position': '2b', 'fileName': 'al-2b+2014.csv', 'data': []},\n",
    "        {'position': '3b', 'fileName': 'al-3b+2014.csv', 'data': []},\n",
    "        {'position': 'ss', 'fileName': 'al-ss+2014.csv', 'data': []},\n",
    "        {'position': 'of', 'fileName': 'al-of+2014.csv', 'data': []},\n",
    "        {'position': 'dh', 'fileName': 'al-dh+2014.csv', 'data': []},\n",
    "        {'position': 'c', 'fileName': 'al-c+2014.csv', 'data': []},\n",
    "    ],\n",
    "    'nl': [\n",
    "        {'position': '1b', 'fileName': 'nl-1b+2014.csv', 'data': []},\n",
    "        {'position': '2b', 'fileName': 'nl-2b+2014.csv', 'data': []},\n",
    "        {'position': '3b', 'fileName': 'nl-3b+2014.csv', 'data': []},\n",
    "        {'position': 'ss', 'fileName': 'nl-ss+2014.csv', 'data': []},\n",
    "        {'position': 'of', 'fileName': 'nl-of+2014.csv', 'data': []},\n",
    "        {'position': 'c', 'fileName': 'nl-c+2014.csv', 'data': []},\n",
    "    ]\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 443,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "winners = {\n",
    "    'al': {\n",
    "        '1b': {\n",
    "            'file': 'al-1b+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Jose Abreu', 'data': [] },\n",
    "                '2015': { 'name': 'Miguel Cabrera', 'data': [] },\n",
    "                '2016': { 'name': 'Miguel Cabrera', 'data': [] },\n",
    "                '2017': { 'name': 'Eric Hosmer', 'data': [] },\n",
    "                '2018': { 'name': 'Jose Abreu', 'data': [] },\n",
    "                '2019': { 'name': 'Carlos Santana', 'data': [] }\n",
    "            }\n",
    "        },\n",
    "        '2b': {\n",
    "            'file': 'al-2b+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Jose Altuve', 'data': [] },\n",
    "                '2015': { 'name': 'Jose Altuve', 'data': [] },\n",
    "                '2016': { 'name': 'Jose Altuve', 'data': [] },\n",
    "                '2017': { 'name': 'Jose Altuve', 'data': [] },\n",
    "                '2018': { 'name': 'Jose Altuve', 'data': [] },\n",
    "                '2019': { 'name': 'DJ LeMahieu', 'data': [] }\n",
    "            }\n",
    "        },\n",
    "        '3b': {\n",
    "            'file': 'al-3b+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Adrian Beltre', 'data': [] },\n",
    "                '2015': { 'name': 'Josh Donaldson', 'data': [] },\n",
    "                '2016': { 'name': 'Josh Donaldson', 'data': [] },\n",
    "                '2017': { 'name': 'Jose Ramirez', 'data': [] },\n",
    "                '2018': { 'name': 'Jose Ramirez', 'data': [] },\n",
    "                '2019': { 'name': 'Alex Bregman', 'data': [] }\n",
    "            }\n",
    "        },\n",
    "        'ss': {\n",
    "            'file': 'al-ss+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Alexei Ramirez', 'data': [] },\n",
    "                '2015': { 'name': 'Xander Bogaerts', 'data': [] },\n",
    "                '2016': { 'name': 'Xander Bogaerts', 'data': [] },\n",
    "                '2017': { 'name': 'Francisco Lindor', 'data': [] },\n",
    "                '2018': { 'name': 'Francisco Lindor', 'data': [] },\n",
    "                '2019': { 'name': 'Xander Bogaerts', 'data': [] }\n",
    "\n",
    "            }\n",
    "        },\n",
    "        'ofs': {\n",
    "            'file': 'al-of+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': {\n",
    "                    'names': ['Jose Bautista', 'Mike Trout', 'Michael Brantley'],\n",
    "                    'data': []\n",
    "                },\n",
    "                '2015': {\n",
    "                    'names': ['Nelson Cruz', 'Mike Trout', 'J.D. Martinez'],\n",
    "                    'data': []\n",
    "                },\n",
    "                '2016': {\n",
    "                    'names': ['Mookie Betts', 'Mike Trout', 'Mark Trumbo'],\n",
    "                    'data': []\n",
    "                },\n",
    "                '2017': {\n",
    "                    'names': ['Aaron Judge', 'George Springer', 'Justin Upton'],\n",
    "                    'data': []\n",
    "                },\n",
    "                '2018': {\n",
    "                    'names': ['Mookie Betts', 'Mike Trout', 'J.D. Martinez'],\n",
    "                    'data': []\n",
    "                },\n",
    "                '2019': {\n",
    "                    'names': ['Mookie Betts', 'Mike Trout', 'George Springer'],\n",
    "                    'data': []\n",
    "                }\n",
    "            }\n",
    "        },\n",
    "        'of1': {\n",
    "            'file': 'al-of+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Jose Bautista', 'data': [] },\n",
    "                '2015': { 'name': 'Nelson Cruz', 'data': [] },\n",
    "                '2016': { 'name': 'Mookie Betts', 'data': [] },\n",
    "                '2017': { 'name': 'Aaron Judge', 'data': [] },\n",
    "                '2018': { 'name': 'Mookie Betts', 'data': [] },\n",
    "                '2019': { 'name': 'Mookie Betts', 'data': [] }\n",
    "            }\n",
    "        },\n",
    "        'of2': {\n",
    "            'file': 'al-of+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Mike Trout', 'data': [] },\n",
    "                '2015': { 'name': 'Mike Trout', 'data': [] },\n",
    "                '2016': { 'name': 'Mike Trout', 'data': [] },\n",
    "                '2017': { 'name': 'George Springer', 'data': [] },\n",
    "                '2018': { 'name': 'Mike Trout', 'data': [] },\n",
    "                '2019': { 'name': 'Mike Trout', 'data': [] }\n",
    "            }\n",
    "        },\n",
    "        'of3': {\n",
    "            'file': 'al-of+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Michael Brantley', 'data': [] },\n",
    "                '2015': { 'name': 'J.D. Martinez', 'data': [] },\n",
    "                '2016': { 'name': 'Mark Trumbo', 'data': [] },\n",
    "                '2017': { 'name': 'Justin Upton', 'data': [] },\n",
    "                '2018': { 'name': 'J.D. Martinez', 'data': [] },\n",
    "                '2019': { 'name': 'George Springer', 'data': [] }\n",
    "\n",
    "            }\n",
    "        },\n",
    "        'c': {\n",
    "            'file': 'al-c+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Yan Gomes', 'data': [] },\n",
    "                '2015': { 'name': 'Brian McCann', 'data': [] },\n",
    "                '2016': { 'name': 'Salvador Perez', 'data': [] },\n",
    "                '2017': { 'name': 'Gary Sanchez', 'data': [] },\n",
    "                '2018': { 'name': 'Salvador Perez', 'data': [] },\n",
    "                '2019': { 'name': 'Mitch Garver', 'data': [] }\n",
    "            }\n",
    "        },\n",
    "        'dh': {\n",
    "            'file': 'al-dh+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Victor Martinez', 'data': [] },\n",
    "                '2015': { 'name': 'Kendrys Morales', 'data': [] },\n",
    "                '2016': { 'name': 'David Ortiz', 'data': [] },\n",
    "                '2017': { 'name': 'Nelson Cruz', 'data': [] },\n",
    "                '2018': { 'name': 'J.D. Martinez', 'data': [] },\n",
    "                '2019': { 'name': 'Nelson Cruz', 'data': [] }\n",
    "            }\n",
    "        }\n",
    "    },\n",
    "    'nl': {\n",
    "        '1b': {\n",
    "            'file': 'nl-1b+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Adrian Gonzalez', 'data': [] },\n",
    "                '2015': { 'name': 'Paul Goldschmidt', 'data': [] },\n",
    "                '2016': { 'name': 'Anthony Rizzo', 'data': [] },\n",
    "                '2017': { 'name': 'Paul Goldschmidt', 'data': [] },\n",
    "                '2018': { 'name': 'Paul Goldschmidt', 'data': [] },\n",
    "                '2019': { 'name': 'Freddie Freeman', 'data': [] }\n",
    "            }\n",
    "        },\n",
    "        '2b': {\n",
    "            'file': 'nl-2b+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Neil Walker', 'data': [] },\n",
    "                '2015': { 'name': 'Dee Gordon', 'data': [] },\n",
    "                '2016': { 'name': 'Daniel Murphy', 'data': [] },\n",
    "                '2017': { 'name': 'Daniel Murphy', 'data': [] },\n",
    "                '2018': { 'name': 'Javier Baez', 'data': [] },\n",
    "                '2019': { 'name': 'Ozzie Albies', 'data': [] }\n",
    "            }\n",
    "        },\n",
    "        '3b': {\n",
    "            'file': 'nl-3b+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Anthony Rendon', 'data': [] },\n",
    "                '2015': { 'name': 'Nolan Arenado', 'data': [] },\n",
    "                '2016': { 'name': 'Nolan Arenado', 'data': [] },\n",
    "                '2017': { 'name': 'Nolan Arenado', 'data': [] },\n",
    "                '2018': { 'name': 'Nolan Arenado', 'data': [] },\n",
    "                '2019': { 'name': 'Anthony Rendon', 'data': [] }\n",
    "            }\n",
    "        },\n",
    "        'ss': {\n",
    "            'file': 'nl-ss+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Ian Desmond', 'data': [] },\n",
    "                '2015': { 'name': 'Brandom Crawford', 'data': [] },\n",
    "                '2016': { 'name': 'Corey Seager', 'data': [] },\n",
    "                '2017': { 'name': 'Corey Seager', 'data': [] },\n",
    "                '2018': { 'name': 'Trevor Story', 'data': [] },\n",
    "                '2019': { 'name': 'Trevor Story', 'data': [] }\n",
    "            }\n",
    "        },\n",
    "        'ofs': {\n",
    "            'file': 'nl-of+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': {\n",
    "                    'names': ['Andrew McCutchen', 'Giancarlo Stanton', 'Justin Upton'],\n",
    "                    'data': []\n",
    "                },\n",
    "                '2015': {\n",
    "                    'names': ['Andrew McCutchen', 'Bryce Harper', 'Carlos Gonzalez'],\n",
    "                    'data': []\n",
    "                },\n",
    "                '2016': {\n",
    "                    'names': ['Christian Yelich', 'Yoenis Cespedes', 'Charlie Blackmon'],\n",
    "                    'data': []\n",
    "                },\n",
    "                '2017': {\n",
    "                    'names': ['Marcell Ozuna', 'Giancarlo Stanton', 'Charlie Blackmon'],\n",
    "                    'data': []\n",
    "                },\n",
    "                '2018': {\n",
    "                    'names': ['Christian Yelich', 'Nick Markakis', 'David Peralta'],\n",
    "                    'data': []\n",
    "                },\n",
    "                '2019': {\n",
    "                    'names': ['Cody Bellinger', 'Christian Yelich', 'Ronald Acuna Jr.'],\n",
    "                    'data': []\n",
    "                }\n",
    "            }\n",
    "        },\n",
    "        'of1': {\n",
    "            'file': 'nl-of+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Andrew McCutchen', 'data': [] },\n",
    "                '2015': { 'name': 'Andrew McCutchen', 'data': [] },\n",
    "                '2016': { 'name': 'Christian Yelich', 'data': [] },\n",
    "                '2017': { 'name': 'Marcell Ozuna', 'data': [] },\n",
    "                '2018': { 'name': 'Christian Yelich', 'data': [] },\n",
    "                '2019': { 'name': 'Cody Bellinger', 'data': [] }\n",
    "            }\n",
    "        },\n",
    "        'of2': {\n",
    "            'file': 'nl-of+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Giancarlo Stanton', 'data': [] },\n",
    "                '2015': { 'name': 'Bryce Harper', 'data': [] },\n",
    "                '2016': { 'name': 'Yoenis Cespedes', 'data': [] },\n",
    "                '2017': { 'name': 'Giancarlo Stanton', 'data': [] },\n",
    "                '2018': { 'name': 'Nick Markakis', 'data': [] },\n",
    "                '2019': { 'name': 'Christian Yelich', 'data': [] }\n",
    "            }\n",
    "        },\n",
    "        'of3': {\n",
    "            'file': 'nl-of+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Justin Upton', 'data': [] },\n",
    "                '2015': { 'name': 'Carlos Gonzalez', 'data': [] },\n",
    "                '2016': { 'name': 'Charlie Blackmon', 'data': [] },\n",
    "                '2017': { 'name': 'Charlie Blackmon', 'data': [] },\n",
    "                '2018': { 'name': 'David Peralta', 'data': [] },\n",
    "                '2019': { 'name': 'Ronald Acuna Jr.', 'data': [] }\n",
    "            }\n",
    "        },\n",
    "        'c': {\n",
    "            'file': 'nl-c+2014.csv',\n",
    "            'stats': {\n",
    "                '2014': { 'name': 'Buster Posey', 'data': [] },\n",
    "                '2015': { 'name': 'Buster Posey', 'data': [] },\n",
    "                '2016': { 'name': 'Wilson Ramos', 'data': [] },\n",
    "                '2017': { 'name': 'Buster Posey', 'data': [] },\n",
    "                '2018': { 'name': 'J.T. Realmuto', 'data': [] },\n",
    "                '2019': { 'name': 'J.T. Realmuto', 'data': [] }\n",
    "            }\n",
    "        }\n",
    "    }\n",
    "}\n",
    "predWinners = {}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 444,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def recall(value):\n",
    "    counter = 0\n",
    "    for league in winners:\n",
    "        for position in winners[league]:\n",
    "            if position == 'of1' or position == 'of2' or position == 'of3' or position == 'of':\n",
    "                continue\n",
    "            if position == 'ofs':\n",
    "                for of in ['of1', 'of2', 'of3']:\n",
    "                    try:\n",
    "                        index_elem = predWinners[league][position].index(winners[league][of]['stats']['2019']['name'])\n",
    "                        if (index_elem+1) <= value * 3:\n",
    "#                             print(index_elem, winners[league][of]['stats'][str(lastYear + 1)]['name'])\n",
    "                            counter = counter + 1\n",
    "                    except ValueError:\n",
    "                        continue\n",
    "            else:\n",
    "                try:\n",
    "                    index_elem = predWinners[league][position].index(winners[league][position]['stats']['2019']['name'])\n",
    "                    if (index_elem+1) <= value:\n",
    "#                         print(index_elem, winners[league][position]['stats']['2019']['name'])\n",
    "                        counter = counter + 1\n",
    "                except ValueError:\n",
    "                    continue\n",
    "    return counter   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 445,
   "metadata": {},
   "outputs": [],
   "source": [
    "def ranking():\n",
    "    accum = 0\n",
    "    for league in winners:\n",
    "        for position in winners[league]:\n",
    "            if position == 'of1' or position == 'of2' or position == 'of3' or position == 'of':\n",
    "                continue\n",
    "            if position == 'ofs':\n",
    "                for of in ['of1', 'of2', 'of3']:\n",
    "                    try:\n",
    "                        index_elem = predWinners[league][position].index(winners[league][of]['stats'][str(lastYear + 1)]['name'])\n",
    "#                         print(index_elem, winners[league][of]['stats']['2019']['name'])\n",
    "                        accum = accum + 1 / math.ceil(((index_elem + 1) / 3))\n",
    "                    except ValueError:\n",
    "                        continue\n",
    "            else:\n",
    "                try:\n",
    "                    index_elem = predWinners[league][position].index(winners[league][position]['stats'][str(lastYear + 1)]['name'])\n",
    "#                     print(index_elem, winners[league][position]['stats']['2019']['name'])\n",
    "                    accum = accum + 1 / (index_elem + 1)\n",
    "                except ValueError:\n",
    "                    continue\n",
    "    return accum"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 446,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def printStats(mode):\n",
    "    if mode == 'predicted':\n",
    "        colList = ['Name', 'runsPredicted', 'HRsPredicted', 'RBIsPredicted', 'AVGPredicted', 'opsPredicted']\n",
    "    elif mode == 'linearSVR':\n",
    "        colList = ['Name', 'linearSVRRuns', 'linearSvrHR', 'linearSvrRBI', 'linearSvrAVG', 'linearSvrOPS']\n",
    "    elif mode == 'svr':\n",
    "        colList = ['Name', 'svrRuns', 'SvrHR', 'svrRBI', 'svrAVG', 'svrOPS']\n",
    "    elif mode == 'huber':\n",
    "        colList = ['Name', 'HuberRegressorRBI', 'HuberRegressorAVG', 'HuberRegressorHR', 'HuberRegressorOPS', 'HuberRegressorRuns']\n",
    "    elif mode == 'ridge':\n",
    "        colList = ['Name', 'RidgeRBI', 'RidgeAVG', 'RidgeHR', 'RidgeOPS', 'RidgeRuns']\n",
    "    else:\n",
    "        if lastYear == 2018 or lastYear == 2019:\n",
    "            colList = ['Name', 'runsActual', 'HRsActual', 'RBIsActuals', 'AVGActuals', 'OpsActuals']\n",
    "        else:\n",
    "            colList = ['Name', 'runs', 'hr', 'rbi', 'avg', 'ops']\n",
    "    for league in predWinners:\n",
    "        for position in predWinners[league]:\n",
    "            if position == 'ofs':\n",
    "                for  i in range(0, 3):\n",
    "                    winner = predWinners[league][position][i]\n",
    "                    winnerStats = allData.loc[allData['Name'] == winner]\n",
    "                    winnerStats = winnerStats[colList]\n",
    "                    print(winnerStats)\n",
    "            else:\n",
    "                winner = predWinners[league][position][0]\n",
    "                winnerStats = allData.loc[allData['Name'] == winner]\n",
    "                winnerStats = winnerStats[colList]\n",
    "                print(winnerStats)\n",
    "            "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 447,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2019\n",
      "5\n",
      "7\n",
      "11\n",
      "11\n",
      "13\n",
      "7.733333333333333\n"
     ]
    }
   ],
   "source": [
    "# divide players by league and by position\n",
    "mode = 'ridge'\n",
    "dataMode = 'scale'\n",
    "neighbors = 1\n",
    "for league in specificData:\n",
    "    for position in specificData[league]:\n",
    "        position['data'] = getDataByLAP(position['fileName'], mode, dataMode)\n",
    "compute_by_position(dataMode)\n",
    "# winners['al']['1b']\n",
    "print(lastYear+1)\n",
    "knn(mode, neighbors, False)\n",
    "# printStats(mode) # uncomment this when you need to see the predicted winner stats\n",
    "for i in range (1, 6):\n",
    "    val = recall(i)\n",
    "    print(val)\n",
    "print(ranking())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "combinedWinners = pd.concat([winners['al']['1b']['average'], winners['nl']['1b']['average']])\n",
    "combinedWinners = combinedWinners.mean(axis=0)"
   ]
  }
 ],
 "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}