In [23]:

import pandas as pd
import numpy as np
from tqdm.notebook import tqdm
import matplotlib.pyplot as plt

%matplotlib inline

In [24]:

from xgboost import XGBClassifier
from sklearn.metrics import brier_score_loss, roc_auc_score
from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.calibration import CalibratedClassifierCV
from scikitplot.metrics import plot_calibration_curve

In [25]:

import warnings
warnings.filterwarnings('ignore', category=pd.io.pytables.PerformanceWarning)

In [26]:

import os
data_dir = os.getcwd() + '/data/wy_scout/'

In [27]:

df_games = pd.read_hdf(data_dir + 'spadl.h5', 'games')
df_games.tail(10)

Out[27]:

	game_id	competition_id	season_id	game_date	home_team_id	away_team_id
54	2057984	28	10078	2018-06-17 15:00:00	3148	15473
55	2057979	28	10078	2018-06-17 12:00:00	16871	17322
56	2057973	28	10078	2018-06-16 19:00:00	9598	16823
57	2057967	28	10078	2018-06-16 16:00:00	15594	7712
58	2057972	28	10078	2018-06-16 13:00:00	12274	7839
59	2057966	28	10078	2018-06-16 10:00:00	4418	8493
60	2057960	28	10078	2018-06-15 18:00:00	9905	1598
61	2057961	28	10078	2018-06-15 15:00:00	16216	10840
62	2057955	28	10078	2018-06-15 12:00:00	16129	15670
63	2057954	28	10078	2018-06-14 15:00:00	14358	16521

In [28]:

dfs_features = []
for _, game in tqdm(df_games.iterrows(), total=len(df_games)):
    game_id = game['game_id']
    df_features = pd.read_hdf(data_dir + 'features.h5', key=f'game_{game_id}')
    df_features['game_id'] = game_id
    dfs_features.append(df_features)
df_features = pd.concat(dfs_features).reset_index(drop=True)
df_features.tail(10)

HBox(children=(FloatProgress(value=0.0, max=64.0), HTML(value='')))

Out[28]:

	type_pass_a0	type_cross_a0	type_throw_in_a0	type_freekick_crossed_a0	type_freekick_short_a0	type_corner_crossed_a0	type_corner_short_a0	type_take_on_a0	type_foul_a0	type_tackle_a0	...	end_angle_to_goal_a0	end_dist_to_goal_a1	end_angle_to_goal_a1	end_dist_to_goal_a2	end_angle_to_goal_a2	team_1	team_2	time_delta_1	time_delta_2	game_id
80969	False	False	False	False	False	False	False	False	False	True	...	0.088083	46.801607	0.160512	38.019553	0.107520	True	True	3.724954	5.269374	2057954
80970	True	False	False	False	False	False	False	False	False	False	...	0.313155	46.379806	0.088083	46.801607	0.160512	True	True	1.175848	4.900802	2057954
80971	True	False	False	False	False	False	False	False	False	False	...	0.163867	37.524973	0.313155	46.379806	0.088083	True	True	1.412850	2.588698	2057954
80972	False	False	False	False	False	False	False	False	False	False	...	0.284537	50.020077	0.163867	37.524973	0.313155	True	True	1.416485	2.829335	2057954
80973	True	False	False	False	False	False	False	False	False	False	...	0.485553	33.913609	0.284537	50.020077	0.163867	True	True	1.416484	2.832969	2057954
80974	True	False	False	False	False	False	False	False	False	False	...	0.532102	32.055017	0.485553	33.913609	0.284537	True	True	2.531683	3.948167	2057954
80975	False	False	False	False	False	False	False	False	False	True	...	0.108906	83.021503	0.164555	78.096249	0.192750	False	False	2.483783	5.015466	2057954
80976	False	False	False	False	False	False	False	False	True	False	...	0.108906	81.331839	0.108906	83.021503	0.164555	True	False	1.862786	4.346569	2057954
80977	False	False	False	False	False	False	False	False	False	False	...	1.570796	25.717078	0.350897	25.717078	0.350897	False	False	66.594979	68.457765	2057954
80978	True	False	False	False	False	False	False	False	False	False	...	1.570796	105.055033	0.032370	81.331839	0.108906	False	True	43.943937	110.538916	2057954

10 rows × 143 columns

In [29]:

dfs_labels = []
"""
for _, game in tqdm(df_games.iterrows(), total=len(df_games)):
    game_id = game['game_id']
    df_labels = pd.read_hdf()
"""

Out[29]:

"\nfor _, game in tqdm(df_games.iterrows(), total=len(df_games)):\n    game_id = game['game_id']\n    df_labels = pd.read_hdf()\n"