-
Notifications
You must be signed in to change notification settings - Fork 0
/
predictFPL_app.py
310 lines (251 loc) · 9.51 KB
/
predictFPL_app.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
import pandas as pd
import re
import random
import requests
import pickle
from tqdm.auto import tqdm
import panel as pn
# set url for fantasy PL API
api_url = "https://fantasy.premierleague.com/api/bootstrap-static/"
# download the webpage
data = requests.get(api_url)
json = data.json()
json.keys()
# build a dataframe
players = pd.DataFrame(json['elements'])
# use all columns
players_df_select = players
# combine first and last names to get player full names
players_df_select['full_name'] = players_df_select[['first_name', 'second_name']].agg(' '.join, axis=1)
# drop first and last name columns
players_df_select = players_df_select.drop(['first_name', 'second_name'], axis = 1)
# player prices are 10x the true value. Divide the prices by 10 to get the true values
players_df_select['now_cost'] = players_df_select['now_cost']/10
# get team info
teams = pd.DataFrame(json['teams'])
# get team defensive strength
team_strength_def = teams[['id', 'name', 'strength_defence_away', 'strength_defence_home']]
# get team attack strength
team_strength_att = teams[['id', 'name', 'strength_attack_away', 'strength_attack_home']]
# get position information from 'element_types'
positions = pd.DataFrame(json['element_types'])
# merge player data with teams and positions
player_team_merge = pd.merge(
left = players_df_select,
right = teams,
left_on = 'team',
right_on = 'id'
)
# merge players with positions
player_team_pos_merge = pd.merge(
left = player_team_merge,
right = positions,
left_on = 'element_type',
right_on = 'id'
)
# rename columns
player_team_pos_merge = player_team_pos_merge.rename(
columns={'name':'team_name', 'singular_name_short':'position_name'}
)
# function for getting specific player gameweek history
def get_history(player_id):
''' get all gameweek history for a given player'''
# request data from API
data = requests.get("https://fantasy.premierleague.com/api/element-summary/" + str(player_id) + "/")
json = data.json()
# turn data into Pandas dataframe
df = pd.DataFrame(json['history'])
return df
tqdm.pandas()
# join team name
players = players.merge(
teams[['id', 'name']],
left_on='team',
right_on='id',
suffixes=['_player', None]
).drop(['team', 'id'], axis=1)
# join player positions
players = players.merge(
positions[['id', 'singular_name_short']],
left_on='element_type',
right_on='id'
).drop(['element_type', 'id'], axis=1)
# rename columns
players = players.rename(
columns={'name':'team', 'singular_name_short':'position'}
)
# get gameweek history for all players
points = players['id_player'].progress_apply(get_history)
# combine results into one dataframe
points = pd.concat(df for df in points)
# join full_name
points = players[['id_player', 'full_name', 'team', 'position']].merge(
points,
left_on='id_player',
right_on='element'
)
# merge opponent defensive strength
points = pd.merge(left = points,
right = team_strength_def[['id', 'strength_defence_away',
'strength_defence_home']],
how = 'left',
left_on = 'opponent_team',
right_on = 'id'
).drop(
'id', axis = 1
).rename(
columns={'strength_defence_away':'opp_def_strength_away', 'strength_defence_home':'opp_def_strength_home'}
)
# assign correct home/away opponent defensive strength for each fixture
def opp_def_strength(row):
if row['was_home'] == False:
return row['opp_def_strength_home']
elif row['was_home'] == True:
return row['opp_def_strength_away']
else:
return "Unknown"
points['opp_def_strength'] = points.apply(lambda row: opp_def_strength(row), axis = 1)
points = points.drop(['opp_def_strength_home','opp_def_strength_away'], axis = 1)
# merge opponent attack strength
points = pd.merge(left = points,
right = team_strength_att[['id', 'strength_attack_away',
'strength_attack_home']],
how = 'left',
left_on = 'opponent_team',
right_on = 'id'
).drop(
'id', axis = 1
).rename(
columns={'strength_attack_away':'opp_att_strength_away',
'strength_attack_home':'opp_att_strength_home'}
)
# assign correct home/away opponent attack strength for each fixture
def opp_att_strength(row):
if row['was_home'] == False:
return row['opp_att_strength_home']
elif row['was_home'] == True:
return row['opp_att_strength_away']
else:
return "Unknown"
points['opp_att_strength'] = points.apply(lambda row: opp_att_strength(row), axis = 1)
points = points.drop(['opp_att_strength_home','opp_att_strength_away'], axis = 1)
# get 20 top scoring players in all positions
gks = points.loc[points['position'] == 'GKP']
defs = points.loc[points['position'] == 'DEF']
mids = points.loc[points['position'] == 'MID']
fwds = points.loc[points['position'] == 'FWD']
top_20_gks = gks.groupby(
['element', 'full_name']
).agg(
{'total_points':'sum'}
).reset_index(
).sort_values(
'total_points', ascending=False
).head(20)
top_20_defs = defs.groupby(
['element', 'full_name']
).agg(
{'total_points':'sum'}
).reset_index(
).sort_values(
'total_points', ascending=False
).head(20)
top_20_mids = mids.groupby(
['element', 'full_name']
).agg(
{'total_points':'sum'}
).reset_index(
).sort_values(
'total_points', ascending=False
).head(20)
top_20_fwds = fwds.groupby(
['element', 'full_name']
).agg(
{'total_points':'sum'}
).reset_index(
).sort_values(
'total_points', ascending=False
).head(20)
#combine top 20 scorers
top_20_all_pos = pd.concat([top_20_gks, top_20_defs, top_20_mids, top_20_fwds], axis = 0)
# select columns of interest
points_select = points[['id_player', 'full_name', 'team', 'position',
'total_points',
'minutes', 'goals_scored', 'assists', 'clean_sheets',
'goals_conceded', 'own_goals',
'saves', 'bonus', 'bps', 'influence', 'creativity', 'threat', 'ict_index',
'expected_goals', 'expected_assists', 'expected_goal_involvements',
'expected_goals_conceded', 'opp_att_strength', 'opp_def_strength']]
points_select['influence'].astype(float)
def last_5_player(df, player_id):
'''
get the mean stats for a given player_id over the last 5 fixtures
prior to most recent fixture and the total points from the most
recent fixture.
assume dataframe is sorted from oldest to newest fixtures
'''
df = df[df['id_player'] == player_id]
last_5 = df.tail(5)
d = {'name': last_5['full_name'].iloc[0],
'id': last_5['id_player'].iloc[0],
'team': last_5['team'].iloc[0],
'position': last_5['position'].iloc[0],
'mean_points': last_5['total_points'].mean(),
'mean_minutes': last_5['minutes'].mean(),
'mean_goals_scored': last_5['goals_scored'].mean(),
'mean_assists': last_5['assists'].mean(),
'mean_clean_sheets': last_5['clean_sheets'].mean(),
'mean_goals_conceded': last_5['goals_conceded'].mean(),
'mean_own_goals': last_5['own_goals'].mean(),
'mean_saves': last_5['saves'].mean(),
'mean_bonus': last_5['bonus'].mean(),
'mean_bps': last_5['bps'].mean(),
'mean_influence': last_5['influence'].astype(float).mean(),
'mean_creativity': last_5['creativity'].astype(float).mean(),
'mean_threat': last_5['threat'].astype(float).mean(),
'mean_ict': last_5['ict_index'].astype(float).mean(),
'mean_xg': last_5['expected_goals'].astype(float).mean(),
'mean_xa': last_5['expected_assists'].astype(float).mean(),
'mean_xgi': last_5['expected_goal_involvements'].astype(float).mean(),
'mean_xgc': last_5['expected_goals_conceded'].astype(float).mean(),
'mean_opp_att': last_5['opp_att_strength'].mean(),
'mean_opp_def': last_5['opp_def_strength'].mean()}
last_5_mean = pd.DataFrame(data = d, index = [0])
return last_5_mean
def last_5_all(df):
''' get last mean stats for all players in df over the last 5 fixtures
prior to most recent fixture and the total points from the most
recent fixture.
'''
last_5_all = pd.DataFrame() # empty dataframe
for p in df['id_player'].unique():
player_df = last_5_player(df, p)
last_5_all = pd.concat([last_5_all, player_df])
return last_5_all
# apply function
last_5_df = last_5_all(points_select)
# load pickled model
with open('mid_model_20230112.pkl', 'rb') as file:
mid_model = pickle.load(file)
# apply the model
def predict_points(player_data, test_data, model) -> None:
'''
apply predictive model to the data
'''
prediction = player_data.assign(predicted=model.predict(test_data))
return prediction[['name', 'team', 'position',
'predicted']].sort_values('predicted', ascending=False).head(10).reset_index()
mid_data = last_5_df[last_5_df['position'] == 'MID']
mid_test = mid_data[['mean_ict', 'mean_xgi']]
# apply prediction model
predicted = predict_points(mid_data, mid_test, mid_model)
# create panel dashboard
pn.extension(sizing_mode="stretch_width")
pn.template.FastListTemplate(
site="PredictFPL",
title="Next Gameweek Player Points Prediction",
sidebar=[],
main=[predicted,
top_20_all_pos.sort_values('total_points', ascending=False)[['full_name', 'total_points']].head(10)],
main_max_width="650px"
).servable();