def list_team_names(my_year):
try:
allTeams = Teams(year=my_year)
except:
print("No data for Year %s" % my_year)
return
for team in allTeams:
print("%-22s %3s" % (team.name, team.abbreviation))
def test_invalid_default_year_reverts_to_previous_year(
self, *args, **kwargs):
flexmock(utils) \
.should_receive('_find_year_for_season') \
.and_return(2018)
teams = Teams()
for team in teams:
assert team._year == '2017'
def setup_method(self, *args, **kwargs):
self.results = {
'rank': 6,
'abbreviation': 'KAN',
'name': 'Kansas City Chiefs',
'wins': 10,
'losses': 6,
'win_percentage': .625,
'post_season_result': LOST_WILD_CARD,
'games_played': 16,
'points_for': 415,
'points_against': 339,
'points_difference': 76,
'margin_of_victory': 4.8,
'strength_of_schedule': -1.3,
'simple_rating_system': 3.4,
'offensive_simple_rating_system': 3.8,
'defensive_simple_rating_system': -0.3,
'yards': 6007,
'plays': 985,
'yards_per_play': 6.1,
'turnovers': 11,
'fumbles': 3,
'first_downs': 322,
'pass_completions': 363,
'pass_attempts': 543,
'pass_yards': 4104,
'pass_touchdowns': 26,
'interceptions': 8,
'pass_net_yards_per_attempt': 7.1,
'pass_first_downs': 198,
'rush_attempts': 405,
'rush_yards': 1903,
'rush_touchdowns': 12,
'rush_yards_per_attempt': 4.7,
'rush_first_downs': 95,
'penalties': 118,
'yards_from_penalties': 1044,
'first_downs_from_penalties': 29,
'percent_drives_with_points': 44.9,
'percent_drives_with_turnovers': 6.3,
'points_contributed_by_offense': 115.88
}
self.abbreviations = [
'RAM', 'NWE', 'PHI', 'NOR', 'JAX', 'KAN', 'DET', 'PIT', 'RAV',
'MIN', 'SEA', 'CAR', 'SDG', 'DAL', 'ATL', 'WAS', 'HTX', 'TAM',
'OTI', 'SFO', 'GNB', 'BUF', 'RAI', 'NYJ', 'CRD', 'CIN', 'DEN',
'MIA', 'CHI', 'CLT', 'NYG', 'CLE'
]
flexmock(utils) \
.should_receive('_todays_date') \
.and_return(MockDateTime(YEAR, MONTH))
self.teams = Teams()
def test_nfl_empty_page_returns_no_teams(self):
flexmock(utils) \
.should_receive('_no_data_found') \
.once()
flexmock(utils) \
.should_receive('_get_stats_table') \
.and_return(None)
teams = Teams()
assert len(teams) == 0
def addTeams(conn, startYear, endYear):
teamz = {}
for year in range(startYear, endYear):
seasons = Teams(year=year)
for team in seasons:
if team.abbreviation not in teamz.keys():
sql = addTeamData.format(SCHEMANAME, team.abbreviation,
team.name)
sqhelp.transaction(conn, sql)
teamz[team.abbreviation] = ""
sqhelp.addTeamSeasonStats(conn, team, year)
return teamz
def list_team_names(my_year):
# if you give an unknown year or team it will go to the except block which prints out no data for that year
# then return to the main page in this code
try:
allTeams = Teams(year=my_year)
except:
print("No data for Year %s" % my_year)
return
# walk through each team and print out just the long name and the abbreviated name
for team in allTeams:
print("%-22s %-3s" % (team.name, team.abbreviation))
def test_teams_string_representation(self, *args, **kwargs):
expected = """Los Angeles Rams (RAM)
New England Patriots (NWE)
Philadelphia Eagles (PHI)
New Orleans Saints (NOR)
Jacksonville Jaguars (JAX)
Kansas City Chiefs (KAN)
Detroit Lions (DET)
Pittsburgh Steelers (PIT)
Baltimore Ravens (RAV)
Minnesota Vikings (MIN)
Seattle Seahawks (SEA)
Carolina Panthers (CAR)
Los Angeles Chargers (SDG)
Dallas Cowboys (DAL)
Atlanta Falcons (ATL)
Washington Redskins (WAS)
Houston Texans (HTX)
Tampa Bay Buccaneers (TAM)
Tennessee Titans (OTI)
San Francisco 49ers (SFO)
Green Bay Packers (GNB)
Buffalo Bills (BUF)
Oakland Raiders (RAI)
New York Jets (NYJ)
Arizona Cardinals (CRD)
Cincinnati Bengals (CIN)
Denver Broncos (DEN)
Miami Dolphins (MIA)
Chicago Bears (CHI)
Indianapolis Colts (CLT)
New York Giants (NYG)
Cleveland Browns (CLE)"""
teams = Teams()
assert teams.__repr__() == expected
def pull_data():
df = pd.DataFrame()
teams = []
years = []
for i in range(2002, 2020):
teams = Teams(i)
for team in teams:
df = df.append(team.dataframe)
years.append(i)
df["years"] = years
df
#offense only df
Offense_df = df[[
'years', 'abbreviation', 'yards', 'margin_of_victory',
'offensive_simple_rating_system', 'pass_attempts', 'pass_completions',
'pass_first_downs', 'pass_net_yards_per_attempt', 'pass_touchdowns',
'pass_yards', 'percent_drives_with_turnovers',
'percent_drives_with_points', 'plays', 'points_contributed_by_offense',
'rank', 'rush_attempts', 'rush_yards', 'rush_first_downs',
'rush_yards_per_attempt', 'turnovers', 'win_percentage',
'yards_per_play', 'first_downs'
]]
Offense_df
#defense only df
Defense_df = df[[
'years', 'abbreviation', 'defensive_simple_rating_system',
'interceptions', 'points_against'
]]
Defense_df
#misc df
Misc_df = df[[
'years', 'abbreviation', 'fumbles', 'games_played', 'losses', 'name',
'first_downs_from_penalties', 'penalties', 'points_difference',
'points_for', 'post_season_result', 'simple_rating_system',
'strength_of_schedule', 'win_percentage', 'wins',
'yards_from_penalties'
]]
Misc_df
offense = Offense_df.to_dict('records')
defense = Defense_df.to_dict('records')
misc = Misc_df.to_dict('record')
return offense, defense, misc
def generate_teams(year1, year2, include_interval=True):
all_teams_dict = {}
if include_interval:
for year in range(year1, year2 + 1):
teams = Teams(year)
teams_dict = {}
for team in teams:
abbr = team.abbreviation
teams_dict[abbr] = team
all_teams_dict[year] = teams_dict
else:
teams = Teams(year1)
teams_dict = {}
for team in teams:
abbr = team.abbreviation
teams_dict[abbr] = team
all_teams_dict[year1] = teams_dict
teams = Teams(year2)
teams_dict = {}
for team in teams:
abbr = team.abbreviation
teams_dict[abbr] = team
all_teams_dict[year2] = teams_dict
return all_teams_dict
def update_years():
min_year = input("Beginning year: ")
max_year = input("Ending year: ")
while int(min_year) <= int(max_year):
teams = Teams(min_year)
total_teams = len(teams)
count = 1
print('Starting data collection ... \n')
for team in teams:
if not os.path.exists('./data/year/{0}/{1}'.format(
min_year, team.abbreviation).lower()):
os.makedirs('./data/year/{0}/{1}'.format(
min_year, team.abbreviation).lower())
file_csv = './data/year/{0}/{1}/{0}_{1}.csv'.format(
min_year, team.abbreviation).lower()
file_json = './data/year/{0}/{1}/{0}_{1}.json'.format(
min_year, team.abbreviation).lower()
f = open(file_csv, 'w+')
f.close()
team.schedule.dataframe_extended.to_csv(file_csv)
f = open(file_csv, 'r')
reader = csv.reader(f)
lines = list(reader)
lines[0][0] = 'game_id'
f.close()
f = open(file_csv, 'w')
writer = csv.writer(f)
for line in lines:
writer.writerow(line)
f.close()
utils.csv_to_json(file_csv, file_json)
print('{0}/{1} complete'.format(count, total_teams))
count += 1
min_year = utils.year_plus_one(min_year)
def display_teams(my_year):
try:
allTeams = Teams(year=my_year)
except:
print("No data for Year %s" % my_year)
return
# Let's sort the list of teams by wins in ascending order
# The build-in python sorted function takes two parameters
# - the list "allTeams" which was return from call to Teams above
# - the key, which is a small function (lambda) which just returns what we are sorting on "team.wins"
sortedTeams = sorted(allTeams, key=lambda team: team.wins)
# walk through each team and print out more information
for team in sortedTeams:
print(
"%-20s %-11s Won: %2d Lost: %2d pts+: %-3d pts-: %-3d ptdiff: %-4d"
% (team.name, team.abbreviation, team.wins, team.losses,
team.points_for, team.points_against, team.points_difference))
def nfl_build_season_passing_data(year):
"""Build QB Passing dataset from NFL boxscores, for a specified season"""
# get a list of all the teams that competed in the specified year
print(f"Gathering passing data for {year} NFL season:")
teams = Teams(str(year))
# init empty lists for input and output labels; these lists will be populated as matrices
input_arr = []
label_arr = []
# display progress bar
bar = ChargingBar('Analyzing teams', max=len(teams))
# iterate over each team in the list of teams
for team in teams:
# get the schedule (list of games) for the team
games = team.schedule
# for every game in the schedule, if the game has been played, extract:
# the team's passing stats, and the result of the game
for game in games:
# if there is no (raw) result for the game, skip it (probably has yet to be played)
if not game._result:
continue
# append an input dataset to input_arr list (matrix)
input_arr.append(nfl_build_game_passing_data(game))
# append a set of labels to label_arr list (matrix)
label_arr.append(nfl_build_game_labels(game))
# increment progress bar
bar.next()
# sanity check: lengths of input_arr and label_arr must be the same
assert len(input_arr) == len(
label_arr), "Error: input & label array lengths mismatch!"
# finish and return datasets
bar.finish()
return input_arr, label_arr
from sportsreference.nfl.teams import Teams
year_start = 1999 # starting year
year_end = 2018 # ending year
sport = 'NFL'
while year_start <= year_end:
teams = Teams(year_start)
for team in teams:
zscore = (team.wins - (0.5 * (team.wins + team.losses))) / (
(team.wins + team.losses) * 0.5 * 0.5)**(0.5)
zscoreabs = abs(zscore)
print(
f'{sport},{team.name},{team.wins},{team.losses},{year_start},{zscore},{zscoreabs}'
)
year_start += 1
from sportsreference.nfl.teams import Teams
import pandas as pd
dfs = []
for team in Teams():
schedule = team.schedule
df = schedule.dataframe
df['Team'] = team.abbreviation
dfs.append(df)
nfl_sched_df = pd.concat(dfs)
nfl_sched_df = nfl_sched_df.sort_values(by='week')
sched_name_map = {
'CLT': 'IND',
'CRD': 'ARI',
'GNB': 'GB',
'HTX': 'HOU',
'JAX': 'JAC',
'KAN': 'KC',
'NOR': 'NO',
'NWE': 'NE',
'OTI': 'TEN',
'RAI': 'LV',
'RAM': 'LAR',
'RAV': 'BAL',
'SDG': 'LAC',
'SFO': 'SF',
'TAM': 'TB',
}
from sportsreference.nfl.teams import Teams
from sportsreference.nfl.roster import Roster
from sportsreference.nfl.roster import Player
import numpy as np
import pandas as pd
# In[3]:
team_array = []
teams = Teams()
for team in teams:
team_array.append(team.abbreviation)
print(team_array)
# In[4]:
player_array = []
tm_n = len(team_array)
for yr in range (2008, 2019):
for i in range(tm_n):
try:
roster = Roster(team_array[i], yr)
import time
from sportsreference.nfl.teams import Teams
import pandas as pd
defense_list = []
for year in ['2010', '2011', '2012']:
teams = Teams(year)
for team in teams:
i = 1
for game in team.schedule:
t0 = time.time()
if game.location=='Home':
prefix='away'
else:
prefix='home'
game_df = game.boxscore.dataframe
row = [
team.abbreviation,
game.datetime,
game_df['{}_fumbles_lost'.format(prefix)].values[0],
game_df['{}_interceptions'.format(prefix)].values[0],
game_df['{}_pass_yards'.format(prefix)].values[0],
game_df['{}_rush_yards'.format(prefix)].values[0],
game_df['{}_times_sacked'.format(prefix)].values[0],
game_df['{}_points'.format(prefix)].values[0]
]
defense_list.append(row)
print('{}, {} game {} finished in {:.2f}s'.format(year, team.abbreviation, i, time.time() - t0))
i += 1
def display_results(my_year):
# stores the abbreviated names in a dictionary which converts abbr names to full name
abbrevName2Name = {}
allTeams = Teams(year=my_year)
# walk through all the teams, saving a dictionary of abbrev name to actual team name
for team in allTeams:
abbrevName2Name[team.abbreviation] = team.name
# Loop forever
while 1:
name = input("Team Name (l/list/q/quit/stop): ")
# this time we convert to UPPER, because we need abbreviated name in UpperCase
name = name.upper()
if (name == "") or (name == "Q") or (name == "QUIT") or (name
== "STOP"):
break
if (name == "L") or (name == "LIST"):
list_team_names(my_year)
# "continue" goes back up to run the while loop again
continue
try:
allgames = Schedule(name, year=my_year)
except:
print("This is an unknown team or an unavaliable year")
# "continue" goes back up to run the while loop again
continue
teamName = abbrevName2Name[name]
won = 0
lost = 0
tie = 0
# Walk through all the the games, keeping track of
# - number of wins
# - number of losses
# - number of ties
# And print out each games result to the user
for game in allgames:
# if points_allowed is equal "None", then there are no more valid games in this list
if game.points_allowed is None:
break
oppAbbr = game.opponent_abbr.upper()
oppName = abbrevName2Name[oppAbbr]
if game.result is None:
result = "Not Played"
else:
result = game.result
if game.points_scored > game.points_allowed:
won = won + 1
elif game.points_scored < game.points_allowed:
lost = lost + 1
elif game.points_scored == game.points_allowed:
tie = tie + 1
print("%s %4s vs %24s %2d to %2d (%s)" %
(teamName, result, oppName, game.points_scored,
game.points_allowed, game.type))
print("Record: Wins: %d Loss: %d Ties: %d" % (won, lost, tie))
from sklearn.linear_model import BayesianRidge
from sklearn.ensemble import RandomForestRegressor
from sklearn.neural_network import MLPRegressor
from sklearn.metrics import r2_score
import plotly.express as px
import plotly.graph_objects as go
from plotly.offline import plot
# what years should we evaluate?
start_year = 1998
end_year = 2018
# query for the data
nfl_ = pd.DataFrame()
for y in range(start_year, end_year + 1):
teams = Teams(year=y)
print("---- " + str(y) + " ----")
for t in teams:
print(t.name)
df = t.schedule.dataframe
df["team"] = t.name
nfl_ = pd.concat([nfl_, df], axis=0)
# remove missing values
nfl_ = nfl_.dropna()
# convert time_of_possession into minutes
nfl_["time_of_possession"] = "00:" + nfl_["time_of_possession"]
nfl_["time_of_possession"] = pd.to_datetime(nfl_["time_of_possession"])
minutes = nfl_["time_of_possession"].dt.minute
minutes = minutes + (nfl_["time_of_possession"].dt.second / 60)
