Closed jkeena88 closed 6 years ago
Do you think we could combine this with neutral site flag? 0/1/2 = regular season home/away/neutral 3 = bowl 4 = playoff 5 = championship
Any downside to this or limitation we force upon ourselves?
First, I don't think home/away make sense since every record is both home AND away the way the rest of the data is structured. Plus I can determine that by the home/away team id. But it could work to do 0/1 = regular season non-neutral/neutral.
Second, there's still a small issue in that bowl games (and theoretically playoff games in the future) can happen in a team's home stadium. It might still be more of a neutral site game than usual but I'd like to be able to identify when this happens (which I could do if it was marked as a non-neutral bowl game using 2 separate columns).
Honestly either way is going to be 99.9% accurate and easy to work with, but I'd probably prefer separate columns for neutral and game type.
Got it. What should I do for conference championships? Do those fall under championship? Or is that another value we need.
Ooh good catch. I'd add another value. So like: 0 = regular season 1 = conference championship 2 = bowl 3 = playoffs 4 = national championship
Thoughts about overseas showcases (Dublin, Australia, etc)? Just handle those like any other neutral site?
Yeah I say just mark it neutral site. Unless/until we get into location data then it's all just gray area besides home/away/neutral.
Current plan:
1) Mark all games as regular season 2) CCGs
Any game after Dec. 1st (except Army vs Navy)
ACC 2005-2016
Big 10 after 2011
Big 12 from 1996-2010
Pac 12 after 2011
SEC 1992-2016
AAC 2015-2016
Conf USA 2005-2016
MAC 1997-2016
Mountain West 2013-2016
WAC 1996-1998
Big East (NONE)
Sun Belt (NONE) 3) Bowl Games
Finished updating game type through 2017-18 season. Will need to keep updating every year. There might be a few mistakes since going back 40 years there is some overlap between the day of the month when the regular season ends some years and the day of the month when the bowls start other years. I was pretty careful, but should double check later.
I know you said bowl games aren't flagged in Sports Reference, but I would like to have this included in our data set even if I have to do it manually.
Rather than a flag I think it should actually be an Int for Game Type, e.g.: 1 = regular season 2 = bowl (non playoff/championship) 3 = playoffs (non championship) 4 = championship ...