A utility that hits TeamRankings and gets the matchup probabilities for each team in the NCAA tournament. You will need to hard-code (blech) all of the team IDs for now, I'm lazy

ncaaMatchupProbability.py

'''
Created on Mar 18, 2013

@author: Matt Dodge
'''

from bs4 import BeautifulSoup as BS
from httplib2 import Http
from urllib import urlencode
import json


west = {
    1:624,
    2:426,
    3:574,
    4:600,
    5:431,
    6:397,
    7:633,
    8:637,
    9:553,
    10:601,
    11:516,
    12:699,
    13:496,
    14:672,
    15:385,
    16:492
}

east = {
    1:424,
    2:636,
    3:420,
    4:630,
    5:572,
    6:666,
    7:423,
    8:388,
    9:500,
    10:602,
    11:433,
    12:400,
    13:581,
    14:527,
    15:440,
    16:464
}

south = {
    1:604,
    2:641,
    3:697,
    4:428,
    5:532,
    6:403,
    7:573,
    8:388,
    9:632,
    10:597,
    11:422,
    12:481,
    13:1073,
    14:687,
    15:944,
    16:595
}

midwest = {
    1:411,
    2:391,
    3:425,
    4:407,
    5:598,
    6:413,
    7:554,
    8:575,
    9:606,
    10:409,
    11:621,
    12:405,
    13:586,
    14:457,
    15:569,
    16:548
}


def getTeamProb(team1, team2):
    h = Http()
    resp, content = h.request(
        'http://www.teamrankings.com/ncaa-tournament/matchups/?t1={0}&t2={1}&p=tournament-overview'.format(team1,team2),
        'GET')
    
    bb = BS(content)
    
    tableRows = bb.select("table.big-text tbody tr")
    
    total = 1
    count = 0
    for tableRow in tableRows:
        total += float(tableRow.find('td').findNextSibling(text=None).contents[0][:-1])
        count += 1

    return total/count/100

if __name__ == '__main__':
    l = midwest
    o = dict()
    
    for seed1 in l:
        for seed2 in l:
            if seed2 > seed1:
                print 'Checking {0} against {1}'.format(l[seed1],l[seed2])
                o[str(seed1) + ','+str(seed2)] = getTeamProb(l[seed1], l[seed2])
            
    print json.dumps(o)