def format_data_for_team_avg_evolution(year):
df = get_league(year)
result = None
for team in df.HomeTeam.unique():
team_df = fc.get_team_matches(df, team)
avgs = fc.get_team_avg_pts(df, team)
points = fc.get_team_evolution(df, team)
res = pd.DataFrame({\
"Club":[team for i in range(0,team_df.shape[0])],
"Date":team_df.Date,\
"avg":avgs.values,\
"m_avg_5":fc.get_team_mobile_avg_pts(df,team,n=5).values,
"n":[i for i in range(0,team_df.shape[0])],\
"pts":points.values})
if result is None:
result = res
else:
result = result.append(res)
return result
def get_features_for_all_matches(year):
db = format_data_for_team_avg_evolution(year)
matches = get_league(year)\
[fc.interesting_cols()]
home_feats = merge_matches("HomeTeam", db, matches)
home_feats = \
get_only_feats_columns_renamed(\
"home_",matches,home_feats)
away_feats = merge_matches("AwayTeam", db, matches)
away_feats = \
get_only_feats_columns_renamed(\
"away_",matches,away_feats)
res = pd.concat([matches.reset_index(),\
home_feats.reset_index(),away_feats.reset_index()]\
,axis=1)
res["day"] = res["home_n"]
res = res.drop(columns=["index","Date","HomeTeam","AwayTeam"\
,"home_n","away_n"])
res["target"] = res.apply(lambda x: categorize(x), axis=1)
return res.loc[res.day > 0]
def test():
db = format_data_for_team_avg_evolution(2014)
matches = get_league(2014)
return merge_matches("HomeTeam", db, matches)
def get_team_data_per_match(year):
df = get_league(year)[-55:]
db = format_data_for_team_avg_evolution(year)
def build_db(year):
df = get_league(year)