def build_csv(overwrite):
if not os.path.exists('../data/p2p_assists/'):
os.makedirs('../data/p2p_assists')
print('MAKING DIRECTORY : ' + './data/p2p_assists')
file_path = '../data/p2p_assists/passing.csv'
if (not os.path.isfile(file_path)) or overwrite:
passing_df = pd.DataFrame(
columns=['Year', 'Receiver_Id', 'Receiver_Name', 'Passer_Id', 'Passer_Name', 'Frequency', 'Pass',
'Assists'])
for year in range(2013, 2016):
base_df = get_general_stats(PlayerOrTeam.P, MeasureTypes.BASE, PerModes.TOTAL, get_year_string(year),
SeasonTypes.REG)
for index, receiver in base_df.iterrows():
print(receiver.PLAYER_NAME)
player_passing_df = get_player_passing_dashboard(receiver.PLAYER_ID, get_year_string(year))
if player_passing_df is not None:
for index2, passer in player_passing_df.iterrows():
row = [year, passer.PLAYER_ID, passer.PLAYER_NAME_LAST_FIRST, passer.PASS_TEAMMATE_PLAYER_ID,
passer.PASS_FROM, passer.FREQUENCY, passer.PASS, passer.AST]
passing_df = passing_df.append(pd.Series(row, index=passing_df.columns), ignore_index=True)
passing_df = passing_df.sort_values(by='Assists', ascending=False)
passing_df.to_csv(file_path)
return passing_df
else:
return pd.read_csv(file_path)
def get_data(overwrite_file):
file_path = '../data/offensive_v_defensive_pace/data.csv'
if not os.path.exists('../data/offensive_v_defensive_pace'):
print("MAKING DIRECTORY!")
os.makedirs('../data/offensive_v_defensive_pace')
else:
print("DIRECTORY EXISTS!")
if os.path.isfile(file_path) and not overwrite_file:
print("ALREADY HAVE OFFENSIVE AND DEFENSIVE PACE DATA!")
return pd.read_csv(file_path)
else:
print("NEED TO GET OFFNESIVE AND DEFENSIVE PACE DATA!")
data_df = pd.DataFrame(
columns=['GP', 'YEAR', 'TEAM_ID', 'TEAM_NAME', 'OFF_TOP', 'DEF_TOP', 'ORTG', 'DRTG', 'PACE'])
for year in range(2013, 2016):
season_year = str(year) + "-" + str(year + 1)[2:4]
print('YEAR => ' + season_year)
year_df = pd.DataFrame(
columns=['GP', 'YEAR', 'TEAM_ID', 'TEAM_NAME', 'OFF_TOP', 'DEF_TOP', 'ORTG', 'DRTG', 'PACE'])
# get advanced team data for each year for ORTG, DRTG, and PACE
print('GETTING ADVANCED TEAM DATA')
year_adv_df = get_general_stats("Team", "Advanced", "Totals", season_year, "Regular+Season")
# get sportsVU possession data for each year for Offensive Time of Possession
print('GETTING SPORTSVU POSSESSION DATA')
year_svu_df = get_sports_vu_stats("Team", "Possessions", "Totals", season_year, "Regular+Season", '0')
# defensive possession time, not given explicitly but you can pass opponent id as a parameter and then sum
# the possession time of teams playing against each team to calculate defensive possession time
for index, team in year_adv_df.iterrows():
print('GETTING POSSESSION TIMES FOR ' + team.TEAM_NAME)
opp_id = team.TEAM_ID
offensive_top = year_svu_df.loc[year_svu_df['TEAM_ID'] == opp_id]['TIME_OF_POSS'].values[0]
print('OFFENSIVE TIME OF POSSESSION => ' + str(offensive_top))
games = year_svu_df.loc[year_svu_df['TEAM_ID'] == opp_id]['GP'].values[0]
team_year_svu_df = get_sports_vu_stats("Team", "Possessions", "Totals", season_year, "Regular+Season",
opp_id)
defensive_top = team_year_svu_df['TIME_OF_POSS'].sum()
print('DEFENSIVE TIME OF POSSESSION =>' + str(defensive_top))
year_df = year_df.append(pd.Series(
[games, season_year, opp_id, team.TEAM_NAME, offensive_top, defensive_top, team.OFF_RATING,
team.DEF_RATING, team.PACE],
index=['GP', 'YEAR', 'TEAM_ID', 'TEAM_NAME', 'OFF_TOP', 'DEF_TOP', 'ORTG', 'DRTG', 'PACE']),
ignore_index=True)
data_df = data_df.append(year_df)
data_df['TOTAL_POSSESSIONS'] = data_df['PACE'] * data_df['GP']
data_df['OFF_PACE'] = data_df['OFF_TOP'] / data_df['TOTAL_POSSESSIONS']
data_df['DEF_PACE'] = data_df['DEF_TOP'] / data_df['TOTAL_POSSESSIONS']
data_df['DISPLAY'] = data_df['TEAM_NAME'] + " " + data_df["YEAR"]
data_df.to_csv(file_path)
return data_df
def chart_assists(data_df, year, num_of_players):
players_df = data.get_general_stats(
data.PlayerOrTeam.P, data.MeasureTypes.BASE, data.PerModes.TOTAL, year, data.SeasonTypes.REG
)
players_df = players_df.sort_values(by=["AST"], ascending=False).head(num_of_players)
for i, player in players_df.iterrows():
player_df = data_df[data_df["PLAYER2_ID"] == player.PLAYER_ID]
charts.make_matplot_scatter_shot_chart(
player_df, player.PLAYER_NAME, player.PLAYER_ID, year, data.SeasonTypes.REG, "assist"
)
charts.make_matplot_hexbin_shot_chart(
player_df, player.PLAYER_NAME, player.PLAYER_ID, year, data.SeasonTypes.REG, "assist"
)
charts.make_matplot_kde_shot_chart(
player_df, player.PLAYER_NAME, player.PLAYER_ID, year, data.SeasonTypes.REG, "assist"
)
charts.make_histogram(player_df, player.PLAYER_NAME, year, data.SeasonTypes.REG, "assist")
def get_data(overwrite):
dir_path = "../data/player_data/synergy_similarity"
if not os.path.exists(dir_path):
os.makedirs(dir_path)
file_path = dir_path + "/synergy_freqs.csv"
if (not os.path.isfile(file_path)) or overwrite:
data_df = get_general_stats("player", "Base", "Totals", "2015-16", "Regular+Season")
data_df = data_df[["PLAYER_ID", "PLAYER_NAME"]]
for index, stype in enumerate(SynergyPlayTypes):
stype_df = get_synergy_stats("player", stype, "offensive")
stype_df = stype_df[["PlayerIDSID", "Time"]]
stype_df.columns = ["PLAYER_ID", stype + "_Frequency"]
data_df = data_df.merge(stype_df, on="PLAYER_ID", how="left")
data_df = data_df.fillna(0)
data_df.to_csv(file_path)
return data_df
else:
return pd.read_csv(file_path)
def get_yearly_data(overwrite_file):
file_path = '../data/scoring_consistency/top_scorers.csv'
if not os.path.exists('../data/scoring_consistency'):
print("MAKING DIRECTORY!")
os.makedirs('../data/scoring_consistency')
else:
print("DIRECTORY EXISTS!")
if os.path.isfile(file_path) and not overwrite_file:
print("ALREADY HAVE TOP SCORERS FILE!")
return pd.read_csv(file_path)
else:
print("NEED TO GET TOP SCORERS FILE")
data_df = pd.DataFrame()
for year in range(1996, 2016):
year_string = str(year) + "-" + str(year + 1)[2:4]
print("GETTING STATS FOR: " + year_string)
year_df = get_general_stats("player", "Base", "PerGame", year_string, "Regular+Season")
year_df = year_df[['PLAYER_ID', 'PLAYER_NAME', 'TEAM_ID', 'TEAM_ABBREVIATION', 'GP', 'PTS']]
year_df["YEAR"] = year_string
year_df["YEAR_DISPLAY"] = " '" + str(year+1)[2:4]
year_df = year_df.loc[year_df["GP"] >= 50]
year_df = year_df.loc[year_df["PTS"] >= 25]
last_names = [x[1] for x in year_df["PLAYER_NAME"].str.split().tolist()]
year_df["DISPLAY"] = last_names + year_df["YEAR_DISPLAY"]
if year == 1996:
data_df = year_df
else:
data_df = data_df.append(year_df)
data_df.to_csv(file_path)
return data_df
import MySQLdb
import data_getters as data
import pandas as pd
db = MySQLdb.connect(host="localhost", user="******", passwd="1qw2SQL3er4", db="nbadb")
cur = db.cursor()
df = pd.DataFrame()
for year in range(1996, 2016):
base_df = data.get_general_stats(
data.PlayerOrTeam.P, data.MeasureTypes.BASE, data.PerModes.TOTAL, year, data.SeasonTypes.REG
)
base_df["YEAR"] = data.get_year_string(year)
base_df = base_df[
[
"PLAYER_ID",
"TEAM_ID",
"YEAR",
"AGE",
"PLAYER_NAME",
"TEAM_ABBREVIATION",
"GP",
"MIN",
"AST",
"BLK",
"DREB",
"OREB",
"FGA",
"FGM",
"FG3A",
