def compare_matchup(game, event, team):
gpbp = playbyplay.PlayByPlay(game_id=game[0], headers=headers)
playByPlay = gpbp.get_data_frames()[0]
home, visitor = get_home_and_away_event(event.value, playByPlay)
## If its a home game for the team
if game[1]:
plot_game_fgs(str(event), (team, team_to_color[team], home), (game[2], team_to_color[game[2]], visitor))
else:
plot_game_fgs(str(event), (game[2], team_to_color[game[2]], home), (team, team_to_color[team], visitor))
def get_playbyplay(team_abb=None, season=None, season_segment=None):
# call get_teams method and store as data frame
# filter df based on team_abb argument and collect teamid
nba_teams_df = get_team_info()
team_ids = list(nba_teams_df[nba_teams_df['abbreviation'].isin(team_abb)]['id'])
# collect game_ids for each team
game_id_list = get_game_ids(team_id = team_ids, season = season, season_type = season_segment)
# collect game info for each team
game_info_df = get_game_info(team_id = team_ids, season = season, season_type = season_segment)
# Start an empty dict
pbp_data_dict = {}
# Call PlayByPlay for every game_id. Store in pbp_data_dict
for game_id in game_id_list:
df = playbyplay.PlayByPlay(game_id).get_data_frames()[0]
pbp_data_dict[game_id] = df
sleep(1)
# Get dict keys
dict_keys = pbp_data_dict.keys()
# Create list of pbp data frames
frames = [pbp_data_dict[key] for key in dict_keys]
# Union the data frame in frames
pbp_df = pd.concat(frames, axis=0)
# Convert dtypes
pbp_df['GAME_ID'] = pbp_df['GAME_ID'].astype(str)
pbp_df['EVENTNUM'] = pbp_df['EVENTNUM'].astype(str).astype(int)
pbp_df['EVENTMSGTYPE'] = pbp_df['EVENTMSGTYPE'].astype(str).astype(int)
pbp_df['EVENTMSGACTIONTYPE'] = pbp_df['EVENTMSGACTIONTYPE'].astype(str).astype(int)
pbp_df['PERIOD'] = pbp_df['PERIOD'].astype(str).astype(int)
pbp_df['PCTIMESTRING'] = pbp_df['PCTIMESTRING'].astype(str)
pbp_df['HOMEDESCRIPTION'] = pbp_df['HOMEDESCRIPTION'].astype(str)
pbp_df['NEUTRALDESCRIPTION'] = pbp_df['NEUTRALDESCRIPTION'].astype(str)
pbp_df['VISITORDESCRIPTION'] = pbp_df['VISITORDESCRIPTION'].astype(str)
pbp_df['SCORE'] = pbp_df['SCORE'].astype(str)
pbp_df['SCOREMARGIN'] = pbp_df['SCOREMARGIN'].astype(str) ## can't convert to int without dealing with None values
# Add date to pbp_df
pbp_df = pbp_df.merge(game_info_df[['GAME_ID', 'GAME_DATE']], on=['GAME_ID'], how='inner')
# Add home/away column
# Add column to ID which team is on offense
# Add event message descriptions to df
pbp_df['EVENTMSGTYPE_DESC'] = pbp_df['EVENTMSGTYPE'].apply(event_msg_type_desc)
# Add event action description
#pbp_df['EVENTMSGACTION_DESC'] = pbp.df['EVENTMSGACTIONTYPE'].apply(event_mgs_action_desc)
# Return
return(pbp_df)
def get_play_by_play_stats(quarter, game):
turnovers = 0
three_attempt = 0
three_make = 0
try:
pbp = playbyplay.PlayByPlay(game['GAME_ID'],
start_period=quarter,
end_period=quarter)
except:
print("ERROR")
print(quarter)
print(game)
return (0, 0, 0, 0, 0)
time.sleep(2)
dic = pbp.get_normalized_dict()['PlayByPlay']
if len(dic) == 0:
return (0, 0, 0, 0, 0)
try:
for x in dic:
if x['VISITORDESCRIPTION'] != None:
if 'Turnover' in x['VISITORDESCRIPTION']:
turnovers += 1
elif '3PT' in x['VISITORDESCRIPTION']:
three_attempt += 1
if 'MISS' not in x['VISITORDESCRIPTION']:
three_make += 1
if x['HOMEDESCRIPTION'] != None:
if 'Turnover' in x['HOMEDESCRIPTION']:
turnovers += 1
elif '3PT' in x['HOMEDESCRIPTION']:
three_attempt += 1
if 'MISS' not in x['HOMEDESCRIPTION']:
three_make += 1
i = len(dic) - 1
while dic[i]['SCORE'] == None:
i -= 1
away_final = int(dic[i]['SCORE'].split("-")[0])
home_final = int(dic[i]['SCORE'].split("-")[1])
return (turnovers, three_make, three_attempt, away_final + home_final,
abs(away_final - home_final))
except:
print("ERROR")
print(game)
print(dic)
def get_stats(team_abbr):
'''
This function will return the last game stats of a particlar team.
Parameters:
team_abbr: pass a string with the abbreviation of a particular team
example> get_stats('CHI')
'''
nba_teams = teams.get_teams()
# Select the dictionary for the Pacers, which contains their team ID
nba_team = [team for team in nba_teams if team['abbreviation'] == team_abbr][0]
nba_team_id = nba_team['id']
# Query for the last regular season game where the Pacers were playing
gamefinder = leaguegamefinder.LeagueGameFinder(team_id_nullable=nba_team_id,
season_nullable=Season.default,
season_type_nullable=SeasonType.regular)
games_dict = gamefinder.get_normalized_dict()
games = games_dict['LeagueGameFinderResults']
game = games[0]
game_id = game['GAME_ID']
game_matchup = game['MATCHUP']
print(f'Searching through {len(games)} game(s) for the game_id of {game_id} where {game_matchup}')
# Query for the play by play of that most recent regular season game
df = playbyplay.PlayByPlay(game_id).get_data_frames()[0]
df.head() #just looking at the head of the data
df.to_csv(team_abbr+'_'+str(game_id)+'.csv')
return print('stats have been exported in the working folder!')
def lastleadchange_create_df(season,
cat,
pickle_flag=0,
plot_flag=0,
tocsv_flag=0):
if pickle_flag == 1:
league_gamelog, gameid_all_list = create_cat_list(season, cat)
wl = 'W'
wl_gamelog = league_gamelog[league_gamelog['WL'] == wl]
gameid_list = wl_gamelog['GAME_ID'].tolist()
team_log_list = wl_gamelog['TEAM_NAME'].tolist()
teamabb_log_list = wl_gamelog['TEAM_ABBREVIATION'].tolist()
lastleadchange_period_list = []
lastleadchange_time_list = []
for game in gameid_list:
print(game)
response_pbp = pbp.PlayByPlay(game_id=game)
time.sleep(0.5)
content_pbp = json.loads(response_pbp.get_json())
results_pbp = content_pbp['resultSets'][0]
headers_pbp = results_pbp['headers']
rows_pbp = results_pbp['rowSet']
league_pbp = pd.DataFrame(rows_pbp)
league_pbp.columns = headers_pbp
league_pbp.loc[league_pbp['SCOREMARGIN'] == 'TIE',
'SCOREMARGIN'] = 0
# league_pbp.loc[league_pbp['SCOREMARGIN'] != 'TIE', 'Tied?'] = 'False'
league_pbp = league_pbp[league_pbp['SCOREMARGIN'].notnull()]
league_pbp['SCOREMARGIN'] = league_pbp['SCOREMARGIN'].astype(int)
if league_pbp['SCOREMARGIN'].iat[-1] < 0:
league_pbp.loc[league_pbp['SCOREMARGIN'] > 0,
'Loser Lead or Tie?'] = 'True'
league_pbp.loc[league_pbp['SCOREMARGIN'] == 0,
'Loser Lead or Tie?'] = 'True'
else:
league_pbp.loc[league_pbp['SCOREMARGIN'] < 0,
'Loser Lead or Tie?'] = 'True'
league_pbp.loc[league_pbp['SCOREMARGIN'] == 0,
'Loser Lead or Tie?'] = 'True'
league_pbp.reset_index(drop=True, inplace=True)
lastleadchange_idx = league_pbp[league_pbp['Loser Lead or Tie?'] ==
'True'].last_valid_index()
if lastleadchange_idx is not None and lastleadchange_idx != len(
league_pbp):
lastleadchange_period_list.append(
league_pbp['PERIOD'][lastleadchange_idx + 1])
lastleadchange_time_list.append(
league_pbp['PCTIMESTRING'][lastleadchange_idx + 1])
elif lastleadchange_idx == len(league_pbp):
lastleadchange_period_list.append(
league_pbp['PERIOD'][lastleadchange_idx])
lastleadchange_time_list.append(
league_pbp['PCTIMESTRING'][lastleadchange_idx])
else:
lastleadchange_period_list.append(league_pbp['PERIOD'][0])
lastleadchange_time_list.append(league_pbp['PCTIMESTRING'][0])
lastleadchange_df = pd.DataFrame(
list(
zip(gameid_list, team_log_list, teamabb_log_list,
lastleadchange_period_list, lastleadchange_time_list)),
columns=['GAME_ID', 'TEAM', 'TEAM_ABB', 'PERIOD', 'TIME'])
file_name = './lastleadchange_' + season + '.pkl'
lastleadchange_df.to_pickle(file_name)
else:
file_name = './lastleadchange_' + season + '.pkl'
lastleadchange_df = pd.read_pickle(file_name)
longest_game = 48 + (lastleadchange_df['PERIOD'].max() - 4) * 5
period_list = lastleadchange_df['PERIOD'].tolist()
period_list = [
period * 12 if period <= 4 else period * 5 + 48
for period in period_list
]
time_split = lastleadchange_df['TIME'].str.split(':', expand=True)
lastleadchange_df['MIN'] = time_split[0].astype(int)
lastleadchange_df['SEC'] = time_split[1].astype(int)
lastleadchange_df['ELAPSED TIME'] = period_list - lastleadchange_df[
'MIN'] - lastleadchange_df['SEC'] / 60
team_gb_mean = lastleadchange_df.groupby(['TEAM_ABB']).mean()
team_gb_mean.sort_values(by='ELAPSED TIME', inplace=True, ascending=False)
if plot_flag == 1:
plt.style.use('Solarize_Light2')
name_list, abb_list, primary1_list = team_colors.return_teamcolors_lists(
)
colors_df = pd.DataFrame(list(zip(abb_list, primary1_list)),
columns=['TEAM_ABB', 'COLOR_1'])
team_gb_mean = pd.merge(team_gb_mean,
colors_df,
on=['TEAM_ABB'],
how='left').set_index('TEAM_ABB')
fig_box, ax_box = plt.subplots()
box_team = lastleadchange_df.boxplot(column='ELAPSED TIME',
by='TEAM_ABB',
ax=ax_box,
vert=False)
fig_bar, ax_bar = plt.subplots()
q_labels = ['Q1', 'Q2', 'Q3', 'Q4']
qtime_labels = [0, 12, 24, 36]
plt.xticks(ticks=qtime_labels, labels=q_labels)
plt.yticks(fontsize=6)
plt.barh(y=team_gb_mean.index,
width=team_gb_mean['ELAPSED TIME'],
color=team_gb_mean['COLOR_1'])
LCfont = {'fontname': 'Lucida Console'}
plt.suptitle('if the Jazz are ahead at halftime, maybe give up',
**LCfont,
fontsize=16)
plt.title(
'AVERAGE TIME IN WHICH TEAM TAKES & KEEPS LEAD | 2020-21, through May 10 | @cale_williams',
**LCfont,
fontsize=6,
loc='left')
plt.axvline(x=qtime_labels[1], ls=':', lw=1.5, color='black')
plt.axvline(x=qtime_labels[2], ls=':', lw=1.5, color='black')
plt.axvline(x=qtime_labels[3], ls=':', lw=1.5, color='black')
plt.savefig('lastleadchange_2020-21.jpeg',
dpi=500,
bbox_inches='tight')
plt.show()
# print(lastleadchange_df)
if tocsv_flag == 1:
lastleadchange_df.to_csv('./lastleadchange_2020-21.csv')
team_gb_mean.to_csv('./team_gb_med_2020-21.csv')
return lastleadchange_df
# find team Ids
from nba_api.stats.static import teams
teams = teams.get_teams()
GSW = [x for x in teams if x['full_name'] == 'Golden State Warriors'][0]
GSW_id = GSW['id']
#game_stats_player
#could not get DateFrom to work, just use loop with years if necessary
from nba_api.stats.library.parameters import SeasonAll
from nba_api.stats.endpoints import playergamelog
import pandas as pd
gamelog_bron = playergamelog.PlayerGameLog(player_id='2544', season = '2018')
df_bron_games_2018 = gamelog_bron.get_data_frames()
gamelog_bron_all = playergamelog.PlayerGameLog(player_id='2544', season = SeasonAll.all)
df_bron_games_all = gamelog_bron_all.get_data_frames()
#find games played by a team or player
from nba_api.stats.endpoints import leaguegamefinder
GSW_games = leaguegamefinder.LeagueGameFinder(team_id_nullable=GSW_id).get_data_frames()[0]
# game play by play data
from nba_api.stats.endpoints import playbyplay
pbp = playbyplay.PlayByPlay('0021900429').get_data_frames()[0]
kareem = playercareerstats.PlayerCareerStats(player_id="76003")
kareem.get_data_frames()[0]
### get all games from a team (current season, Atlanta 1610612737)
from nba_api.stats.endpoints import teamgamelog
gamelog = teamgamelog.TeamGameLog(team_id="1610612737")
gamelog.get_data_frames()[0]
## get play by play from desired game (0021900543 ATL vs DEN)
from nba_api.stats.endpoints import playbyplay
gamepbp = playbyplay.PlayByPlay(game_id="0021900543")
gamepbp.get_data_frames()[0].sort_values("EVENTNUM")
## get traditional box score from the same game
from nba_api.stats.endpoints import boxscoretraditionalv2
boxscore = boxscoretraditionalv2.BoxScoreTraditionalV2(game_id="0021900543")
boxscore.get_data_frames()[0]
## based on teamgamelog and box score, we could calculate the FPPM for each player
## but how to link games to players? have to go through team?
from nba_api.stats.endpoints import playergamelog
def getPbpCSV(gameid):
return playbyplay.PlayByPlay(gameid).get_data_frames()[0].to_csv(
'./game_data/playbyplay.csv')
gamefinder = leaguegamefinder.LeagueGameFinder(
team_id_nullable=pacers_id,
season_nullable=Season.default,
season_type_nullable=SeasonType.regular)
games_dict = gamefinder.get_normalized_dict()
games = games_dict['LeagueGameFinderResults']
game = games[0]
game_id = game['GAME_ID']
game_matchup = game['MATCHUP']
print(
f'Searching through {len(games)} game(s) for the game_id of {game_id} where {game_matchup}'
)
# Query for the play by play of that most recent regular season game
from nba_api.stats.endpoints import playbyplay
df = playbyplay.PlayByPlay(game_id).get_data_frames()[0]
df.head() #just looking at the head of the data
from nba_api.stats.endpoints import shotchartdetail, shotchartlineupdetail, playerdashptshots, playercareerstats
shot_df = shotchartdetail.ShotChartDetail(
team_id=mia_id, player_id='1628389').get_data_frames()[0]
player_career = playercareerstats.PlayerCareerStats(
player_id='1628389').get_data_frames()[0]
shotchartlineupdetail.ShotChartLineupDetail()