def net_v_all_png():
global rankings, last_updated
img = io.BytesIO()
if rankings is None or (date.today() - last_updated).days > 0:
rankings = collect_rankings()
last_updated = date.today()
tmp_rankings = prepare_rankings(rankings)
tmp_rankings['Sagarin'] = tmp_rankings['Sagarin_RK']
tmp_rankings['KenPom'] = tmp_rankings['Pomeroy_RK']
tmp_rankings['BPI'] = tmp_rankings['BPI_RK']
tmp_rankings['NET'] = tmp_rankings['NET Rank']
fig = sns.pairplot(tmp_rankings,
y_vars=['NET'],
x_vars=['Sagarin', 'KenPom', 'BPI', 'RPI'],
plot_kws={'color': 'black'})
fig.set(ylim=(0, 380), xlim=(0, 380))
FigureCanvas(fig.fig).print_png(img)
output = make_response(img.getvalue())
output.headers["Content-Disposition"] = "inline"
output.headers["Content-Type"] = "image/png"
output.cache_control.max_age = 60 * 60 * 24
return output
def data_csv():
global rankings, last_updated
si = io.StringIO()
if rankings is None or (date.today() - last_updated).days > 0:
rankings = collect_rankings()
last_updated = date.today()
conf = request.args.get('conf', '')
if conf == 'REPLACE_ME':
return 'Please replace REPLACE_ME with a valid conference name'
tmp_rankings = prepare_rankings(rankings)
sort = request.args.get('sort', '')
if sort == 'KPvNET':
tmp_rankings[
'KP - NET'] = tmp_rankings['Pomeroy_RK'] - tmp_rankings['NET Rank']
tmp_rankings = tmp_rankings.drop('RPI - NET', axis=1)
tmp_rankings = tmp_rankings.sort_values(by='KP - NET', ascending=True)
tmp_rankings.to_csv(si)
output = make_response(si.getvalue())
output.headers[
"Content-Disposition"] = "attachment; filename=rankings_{}.csv".format(
request.args.get('conf', ''))
output.headers["Content-Type"] = "text/csv"
output.cache_control.max_age = 60 * 60 * 24
return output
def correlation_by_conf_png():
global rankings, last_updated
img = io.BytesIO()
if rankings is None or (date.today() - last_updated).days > 0:
rankings = collect_rankings()
last_updated = date.today()
tmp_rankings = prepare_rankings(rankings)
tmp_rankings['Sagarin'] = tmp_rankings['Sagarin_RK']
tmp_rankings['KenPom'] = tmp_rankings['Pomeroy_RK']
tmp_rankings['BPI'] = tmp_rankings['BPI_RK']
tmp_rankings['NET'] = tmp_rankings['NET Rank']
tmp_rankings['Conference'] = tmp_rankings['Conf']
n = request.args.get('n', '8')
if n == '8':
tmp_rankings['Conference'][tmp_rankings['Conference'].isin(
POWER_SEVEN_CONFS) == False] = 'Other'
fig = sns.pairplot(tmp_rankings,
y_vars=['KenPom', 'BPI', 'RPI', 'NET'],
x_vars=['Sagarin', 'KenPom', 'BPI', 'RPI'],
hue='Conference',
hue_order=[
'American', 'ACC', 'Big 12', 'Big East',
'Big Ten', 'Pac-12', 'SEC', 'Other'
],
markers=['p', 's', 'x', 'H', 'o', 'P', 'D', '*'],
plot_kws=dict(alpha=0.8))
elif n == '2':
tmp_rankings['Conference'][tmp_rankings['Conference'].isin(
POWER_SEVEN_CONFS) == False] = 'Other'
tmp_rankings['Conference'][tmp_rankings['Conference'].isin(
POWER_SEVEN_CONFS)] = 'Power 7'
fig = sns.pairplot(tmp_rankings,
y_vars=['KenPom', 'BPI', 'RPI', 'NET'],
x_vars=['Sagarin', 'KenPom', 'BPI', 'RPI'],
hue='Conference',
hue_order=['Power 7', 'Other'],
markers=['s', 'o'],
plot_kws=dict(alpha=0.5, color='black'))
elif n == 'all':
fig = sns.pairplot(tmp_rankings,
y_vars=['KenPom', 'BPI', 'RPI', 'NET'],
x_vars=['Sagarin', 'KenPom', 'BPI', 'RPI'],
hue='Conference',
plot_kws=dict(alpha=0.8))
fig.set(ylim=(0, 380), xlim=(0, 380))
for i, j in zip(*np.triu_indices_from(fig.axes, 1)):
fig.axes[i, j].set_visible(False)
FigureCanvas(fig.fig).print_png(img)
output = make_response(img.getvalue())
output.headers["Content-Disposition"] = "inline"
output.headers["Content-Type"] = "image/png"
output.cache_control.max_age = 60 * 60 * 24
return output
def outliers_json():
global rankings, last_updated
analyzer = Analyzer()
if rankings is None or (date.today() - last_updated).days > 0:
rankings = collect_rankings()
last_updated = date.today()
tmp_rankings = prepare_rankings(rankings)
tmp_rankings = tmp_rankings.drop(['RPI - NET', 'Conf'], axis=1)
corr = tmp_rankings.corr(method='spearman')
outliers = analyzer.get_outliers(tmp_rankings, corr)
return jsonify(outliers)
def data():
global rankings, last_updated
si = io.StringIO()
if rankings is None or (date.today() - last_updated).days > 0:
rankings = collect_rankings()
last_updated = date.today()
tmp_rankings = prepare_rankings(rankings)
tmp_rankings.to_json(si, orient='index')
output = make_response(si.getvalue())
output.headers["Content-Type"] = "application/json"
output.cache_control.max_age = 60 * 60 * 24
return output
def correlation_json():
global rankings, last_updated
si = io.StringIO()
if rankings is None or (date.today() - last_updated).days > 0:
rankings = collect_rankings()
last_updated = date.today()
tmp_rankings = prepare_rankings(rankings)
tmp_rankings = tmp_rankings.drop(['RPI - NET', 'Conf'], axis=1)
tmp_rankings.corr(method='spearman').round(3).to_json(si, orient='index')
output = make_response(si.getvalue())
output.headers["Content-Type"] = "application/json"
output.cache_control.max_age = 60 * 60 * 24
return output
def correlation_txt():
global rankings, last_updated
with open('templates/output/as_html.txt') as f:
template = f.read()
if rankings is None or (date.today() - last_updated).days > 0:
rankings = collect_rankings()
last_updated = date.today()
tmp_rankings = prepare_rankings(rankings)
tmp_rankings = tmp_rankings.drop(['RPI - NET', 'Conf'], axis=1)
corr = tmp_rankings.corr(method='spearman')
img = io.BytesIO()
tmp_rankings['Sagarin'] = tmp_rankings['Sagarin_RK']
tmp_rankings['KenPom'] = tmp_rankings['Pomeroy_RK']
tmp_rankings['BPI'] = tmp_rankings['BPI_RK']
tmp_rankings['NET'] = tmp_rankings['NET Rank']
fig = sns.pairplot(tmp_rankings,
y_vars=['NET'],
x_vars=['Sagarin', 'KenPom', 'BPI', 'RPI'],
plot_kws={'color': 'black'})
fig.set(ylim=(0, 380), xlim=(0, 380))
plt.savefig(img, format='png')
img.seek(0)
imageBase64 = base64.encodebytes(img.getvalue()).decode()
payload = template.format(
date=date.today().strftime('%d %B %Y'),
sagarinVpomeroy=corr['Sagarin_RK']['Pomeroy_RK'].round(3),
sagarinVrpi=corr['Sagarin_RK']['RPI'].round(3),
sagarinVbpi=corr['Sagarin_RK']['BPI_RK'].round(3),
sagarinVnet=corr['Sagarin_RK']['NET Rank'].round(3),
pomeroyVrpi=corr['Pomeroy_RK']['RPI'].round(3),
pomeroyVbpi=corr['Pomeroy_RK']['BPI_RK'].round(3),
pomeroyVnet=corr['Pomeroy_RK']['NET Rank'].round(3),
rpiVbpi=corr['RPI']['BPI_RK'].round(3),
rpiVnet=corr['RPI']['NET Rank'].round(3),
bpiVnet=corr['BPI_RK']['NET Rank'].round(3),
imageBase64=imageBase64)
output = make_response(payload)
output.headers["Content-Disposition"] = "attachment; filename=corr.txt"
output.headers["Content-Type"] = "text/plain; charset=utf-8"
output.cache_control.max_age = 60 * 60 * 24
return output
def correlation_csv():
global rankings, last_updated
si = io.StringIO()
if rankings is None or (date.today() - last_updated).days > 0:
rankings = collect_rankings()
last_updated = date.today()
conf = request.args.get('conf', '')
if conf == 'REPLACE_ME':
return 'Please replace REPLACE_ME with a valid conference name'
tmp_rankings = prepare_rankings(rankings)
tmp_rankings = tmp_rankings.drop(['RPI - NET', 'Conf'], axis=1)
tmp_rankings.corr(method='spearman').to_csv(si)
output = make_response(si.getvalue())
output.headers[
"Content-Disposition"] = "attachment; filename=correlation_{}.csv".format(
request.args.get('conf', ''))
output.headers["Content-Type"] = "text/csv"
output.cache_control.max_age = 60 * 60 * 24
return output
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
from datetime import date
from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
from flask import render_template, Blueprint, make_response, send_file, jsonify, request
from utils.analyzer import Analyzer
plt.rcParams['figure.dpi'] = 200
hello_blueprint = Blueprint('hello', __name__)
from templates.hello.collect_data import collect_rankings
rankings = collect_rankings()
last_updated = date.today()
POWER_SEVEN_CONFS = [
'Big Ten', 'Big 12', 'Pac-12', 'ACC', 'American', 'SEC', 'Big East'
]
def prepare_rankings(all_ranks):
conf_ranks = all_ranks.groupby('Conf')
conf_ranks = {conf: df for conf, df in conf_ranks}
conf = request.args.get('conf')
if conf is None:
return all_ranks.copy()
if conf == 'PowerSeven':
teams = pd.concat([conf_ranks[team] for team in POWER_SEVEN_CONFS])