def get_single_factor(df, percentile):
list_of_dicts = list()
for team in unique_teams:
df_ = df_helper.query_team(df, team)
x = df_['TotalScore']
# Position of percentile for x dataset
p = np.percentile(x, percentile)
# print("Team: {}, percentile: {}, pos: {}". format(team, percentile, p))
# Filter df, only getting result out of percentile range
percentile_df = filter_by_percentile(df_, percentile, p)
# Create a frecuency dictionary from dataframe
list_of_dicts.append(frecuencies_df(percentile_df, team))
return sum_dicts(list_of_dicts)
def get_weighted_factors(df, percentile):
list_of_dicts = list()
for team in unique_teams:
df_ = df_helper.query_team(df, team)
x = df_['TotalScore']
# Position of percentile for x dataset
p = np.percentile(x, percentile)
opo_p = np.percentile(x, oposite_percentile(percentile))
width = abs(p - opo_p)
# Filter df, only getting result out of percentile range
teams_df = filter_by_percentile(df_, percentile, p)
# Create a frecuency dictionary from dataframe
list_of_dicts.append(weights_df(teams_df, team, p, width))
return sum_dicts(list_of_dicts)
def normal_distribution(df, team, percentile=10):
df_ = df_helper.query_team(df, team)
x = df_['TotalScore']
p_low = np.percentile(x, percentile)
p_med = np.percentile(x, 50)
p_high = np.percentile(x, 100 - percentile)
mean = statistics.mean(x)
sd = statistics.stdev(x)
plt.plot(x, stats.norm.pdf(x, mean, sd))
plt.title('{} normal deviation'.format(team))
xposition = [p_low, p_med, p_high]
for xc in xposition:
plt.axvline(x=xc, color='r', linestyle='-')
plt.show()
def percentile_team(df, team, percentile):
df_ = df_helper.query_team(df, team)
x = df_["TotalScore"]
p = np.percentile(x, percentile)
return p
def mean_team(df, team):
df_ = df_helper.query_team(df, team)
return df_['TotalScore'].mean()