def __init__(self):
conn = establish_db_connection('sqlalchemy').connect()
self.url = "https://kenpom.com/team.php"
self.url_team = lambda x, y: '%s?team=%s&y=%s' % (self.url, str(x),
str(y))
self.df = pd.DataFrame()
teams = pd.read_sql(
'SELECT DISTINCT Team FROM kenpom_season_data WHERE Year = 2019 AND Rank <= 60',
con=conn)
teams = teams['Team'].tolist()
skip_teams = pd.read_sql(
'SELECT DISTINCT Team FROM kenpom_madness_table', con=conn)
skip_teams = skip_teams['Team'].tolist()
for team in skip_teams:
teams.remove(team)
for team in teams:
self._start_session(team)
print "Fetching game results for %s" % team
self.team = team
self._fetch_team_results()
print "Uploading to db..."
self._upload_data()
print ""
print self.df
self.driver.close()
def four_factors_algo(team_a, team_b):
conn = establish_db_connection('sqlalchemy').connect()
team_a = str(team_a)
team_b = str(team_b)
## fetch data
cols = {
'Team': 'team',
'eFG Perc': 'efg_perc',
'Turnover Perc': 'tov_perc',
'Off Rebound Perc': 'orb_perc',
'FTRate': 'ft_rate',
'Opp eFG Perc': 'opp_efg_perc',
'Opp Turnover Perc': 'opp_tov_perc',
'Opp Off Rebound Perc': 'opp_orb_perc',
'Opp FTRate': 'opp_ft_rate'
}
sql_str = 'SELECT `{}` FROM kenpom_four_factors_data WHERE Year = 2019'.format(
'`, `'.join(cols.keys()))
df = pd.read_sql(sql_str, con=conn)
df.rename(index=str, columns=cols, inplace=True)
## clean data
for col in [col for col in df.columns if col != 'team']:
df[col] = df[col] / 100
df = df[df['team'].isin([team_a, team_b])].copy()
## run algo
weights_dict = {'efg': 0.40, 'tov': 0.25, 'orb': 0.20, 'ft_rate': 0.15}
temp = {}
for index, row in df.iterrows():
efg = (row['efg_perc'] - row['opp_efg_perc']) * 100
tov = (row['tov_perc'] - row['opp_tov_perc']) * 100
orb = (row['tov_perc'] - row['opp_tov_perc']) * 100
ft_rate = (row['ft_rate'] - row['opp_ft_rate']) * 100
temp[row['team']] = {
'efg': efg,
'tov': tov,
'orb': orb,
'ft_rate': ft_rate
}
algo_dict = {}
print team_a
for s, w in weights_dict.iteritems():
algo_dict[s] = (temp[team_a].get(s) - temp[team_b].get(s)) * w
spread = round((sum(algo_dict.values()) * 2), 3)
print spread
return spread
def __init__(self, start, end):
self.conn = establish_db_connection('sqlalchemy').connect()
self.dates = range_all_dates(start, end)
self.clean_cols = [
'game_date', 'game_id', 'away_id', 'away_team', 'home_id',
'home_team', 'pts_away', 'pts_home', 'pt_diff_away', 'pts_total',
'win_side', 'win_id'
]
self.clean_scores = pd.DataFrame(columns=self.clean_cols)
self._fetch_raw_scores()
def __init__(self):
self.conn = establish_db_connection('sqlalchemy').connect()
self.games = ['', '', '', '', '', '']
self.stats = [
'EFG_PCT', 'FTA_RATE', 'TM_TOV_PCT', 'OREB_PCT', 'OPP_EFG_PCT',
'OPP_FTA_RATE', 'OPP_TOV_PCT', 'OPP_OREB_PCT'
]
self._fetch_stats()
self._weight_player_stats()
print self._team[self.stats[:] + ['TEAM_ID']]
print self._weighted
def __init__(self):
self.conn = establish_db_connection('sqlalchemy').connect()
self.today = datetime.now().strftime('%Y-%m-%d')
self.prev_day = (datetime.strptime(self.today, '%Y-%m-%d').date() -
timedelta(days=1)).strftime('%Y-%m-%d')
self.games_list = get_games_list(self.prev_day)
self._fetch_game_stats()
self._fetch_final_scores()
self._fetch_spreads()
# self._calculate_results()
self._update_db()
def linear_algo(team_a, team_b):
conn = establish_db_connection('sqlalchemy').connect()
## fetch data
cols = {
'Team': 'team',
'eFG Perc': 'efg_perc',
'Turnover Perc': 'tov_perc',
'Off Rebound Perc': 'orb_perc',
'FTRate': 'ft_rate',
'Opp eFG Perc': 'opp_efg_perc',
'Opp Turnover Perc': 'opp_tov_perc',
'Opp Off Rebound Perc': 'opp_orb_perc',
'Opp FTRate': 'opp_ft_rate'
}
sql_str = 'SELECT `{}` FROM kenpom_four_factors_data WHERE Year = 2019'.format(
'`, `'.join(cols.keys()))
df = pd.read_sql(sql_str, con=conn)
df.rename(index=str, columns=cols, inplace=True)
## adjust percentages for algorithm
for col in [
x for x in df.columns
if x not in ['team', 'opp_ft_rate', 'ft_rate']
]:
df[col] = df[col] / 100
## run algo
weights_dict = {
'efg_perc': 101.54,
'tov_perc': -0.829,
'orb_perc': 0.063,
'ft_rate': 7.57,
'opp_efg_perc': -92.41,
'opp_tov_perc': 1.0428,
'opp_orb_perc': -0.0123,
'opp_ft_rate': -13.095
}
spreads = {}
for team in [team_a, team_b]:
data = df[df['team'] == team].copy()
spread = 0
for s, w in weights_dict.iteritems():
spread += data[s].max() * w
spreads[team] = spread
spread = spreads[team_a] - spreads[team_b]
print spread
return None
def __init__(self, gamedate=None):
self.conn = establish_db_connection('sqlalchemy').connect()
if gamedate is not None:
self.gamedate = gamedate
else:
self.gamedate = datetime.now().date().strftime('%Y-%m-%d')
self.games_df = assets.games_daily(self.gamedate)
self._fetch_agg_stats()
self._determine_b2b()
self._build_predictions()
self._merge_with_spreads()
self._clean_merged_data()
self.merged_df.to_sql('daily_picks',
con=self.conn,
if_exists='replace',
index=False)
self.merged_df.to_sql('historical_picks',
con=self.conn,
if_exists='append',
index=False)
def __init__(self, stats):
self.conn = establish_db_connection('sqlalchemy').connect()
self.stats = stats
def _upload_data(self):
conn = establish_db_connection('sqlalchemy').connect()
self.df.to_sql('kenpom_madness_table',
con=conn,
if_exists='append',
index=None)
def __init__(self):
self.conn = establish_db_connection('sqlalchemy').connect()
import dash
import dash_core_components as dcc
import dash_html_components as html
import plotly.graph_objs as go
from dash.dependencies import Input, Output, State, Event
import dash_table
import styling
import pandas as pd
import madness_helper as helper
from nba_utilities.db_connection_manager import establish_db_connection
app = dash.Dash()
app.title = "Madness"
app.css.append_css({'external_url':"https://stackpath.bootstrapcdn.com/bootstrap/4.2.1/css/bootstrap.min.css"})
conn = establish_db_connection('sqlalchemy').connect()
teams = pd.read_sql('SELECT DISTINCT team FROM kenpom_four_factors_data', con=conn)
teams = [ row['team'] for index, row in teams.iterrows() ]
teams = [ {'label': team, 'value': team} for team in sorted(teams) ]
def serve_layout():
return html.Div(children = [
html.Div(className="app-header", children = [
html.H1('Alge-bracket', className="app-header--title", style=styling.portal_header_txt)], style = styling.portal_header_bgrd),
dcc.Tabs(id="tabs", children=[
#### Head 2 Head Predictor Tab ####
dcc.Tab(label='H2H Predictor', children=[
html.H1(children='Matchup Predictor', style=styling.tab_header),
html.Div(style={'padding':10}),
html.Div(className='row', children=[