sipgo

README

writes odds for bov to CSV

go run types.go utils.go main.go

Collects the moneyline of ~500 events and the score of live games.

benchmarks

at ~20 mbps

• requesting 1564 (431 with NA drop) lines and scores takes ~~~15 mins~~ ~25 seconds
• 1670 (1669 with NA drop) of lines took ~20 seconds

data breakdown

odds:

sport: FOOT, shape: (4, 8)
sport: CRIC, shape: (10, 8)
sport: BASE, shape: (9, 8)
sport: BASK, shape: (70, 8)
sport: BOXI, shape: (35, 8)
sport: TENN, shape: (141, 8)
sport: SOCC, shape: (921, 8)
sport: HAND, shape: (37, 8)
sport: VOLL, shape: (9, 8)
sport: MMA, shape: (63, 8)
sport: FUTS, shape: (1, 8)
sport: TABL, shape: (24, 8)
sport: HCKY, shape: (61, 8)
sport: AURL, shape: (7, 8)
sport: ESPT, shape: (211, 8)
sport: RUGU, shape: (24, 8)
sport: RUGL, shape: (14, 8)
sport: SNOO, shape: (8, 8)
sport: DART, shape: (21, 8)

TODO

• potentially reseparate Row -> Line + Score and write to different db tables (large performance increase)
• n-competitor/tournament event handling (more data)
• prop markets, find good encoding (more data)
• misc sports fix (tennis scores, boxing/mma/cricket odds)

completed

• parallel sport request (limited to 2 competitor games - count: 19) this turned out to have a minimal performance impact on low # of games, which makes sense
• cloud sql upload
• added region country and competition
• added league and round info to Row
• only writes changes
• 3-way ML fix
• scores using semaphores/concurrency

Documentation

Overview

credit montanaflynn/pget.go https://gist.github.com/montanaflynn/ea4b92ed640f790c4b9cee36046a5383 https://medium.com/dev-bits/making-concurrent-http-requests-in-go-programming-language-823b51bb1dc2

for data.nba.com