I’ve added team and player statistics for games in 2006, 2007 and 2008 (example), and will build out individual game stats back to 2001 (example). I’ve fixed several outstanding data issues, including bowl games, overtime games and at the suggestion of Chris McComas, moved conference affiliation from the College model to the CollegeYear model. I’ve added an ever-diminishing list of unbeaten teams.

There is, of course, so much more to do. I’ve got a list of feature ideas a mile long, plus an urgent need for some graphics (using the Google Chart API), team rankings and poll data. But now that the elections are over, I can spend some time working on the site. My long-term goal is to have Fumblerooski be a resource for people seeking to identify that one stat, to study a trend (either on the site or via APIs) or just to lose themselves in college football data. What would you like to see?

• Yearly rosters for teams (example). Thanks to the NCAA for this stuff, which goes back to 2000.
• Player pages (example). These are a little odd in that they represent a player on a team for a single year, so they have links to other seasons for that player. Again, back to 2000.

Coming up: game summary stats for teams, and player game stats.

It’s August, which means that college football is just around the corner. College football is why I don’t volunteer to teach any classes in the fall. It’s why I occasionally compensate my better half for missed Saturday afternoons (although thanks to ESPN 360, I’m not nearly as bad as I was pre-child). So I love college ball, and I love data. That’s where Fumblerooski comes in.

Let me say from the get-go that this is not nearly a finished site. It’s not even halfway there. I’m posting about it now because I’d like to invite people with similar interests to help me build out a site that puts the numbers behind college football front and center. Yes, I have ideas – APIs, for example – but alone Fumblerooski will only ever be so good, and certainly not good enough. That’s why the code behind the site is on github.

The basics: it’s running Django trunk (so, yes, that’s the 1.0 beta candidate right now) and uses MySQL as a backend. Right now I have game results dating back to 1987 for most major schools and spottier coverage for minor ones. In addition, the NCAA releases game-by-game statistics for players and I have some scripts for processing that data, although there’s plenty of room for improvement. Folks who dive into the code may also notice that I started a recruiting dataset as well, but I think that area is well-covered, so it’s not a priority for me at this time. At the moment, Fumblerooski is running on a Joyent 1/2 gig Accelerator with nginx as the Web server.

Most of my work so far as gone towards building out team information. Take my alma mater, Pittsburgh: you can see the results of a given season, check out a series (you can reverse it if you’re one of those WVU fans) or see details of an individual game. The drive chart, which is a fairly recent NCAA addition, is dynamically fetched (and no, I can’t do anything about the colors).

I envision at least two types of contributors: one would help on the coding side with new features (I have plans for aggregate player stuff, but want to wait to see what gets into Django). Another type could be with information: fleshing out coaching details, for example. In my wildest dreams, Fumblerooski gets a severely-needed makeover as well. Any takers? Feel free to sign up at github, or fork the code, or whatever. You can also contact me if you’d like to help in other ways.

Oh, and the name? It was the best football-only term available, but I also got the blessing of Nebraska alum Matt Waite.

I’ve written before about the applicability of database work to sports, particularly local college and professional teams, and there are signs out there that some news organizations are tackling this previously under-served area. With the NFL draft fast approaching, one of those perennial datasets is a team’s draft history – a topic of discussion among fans, although not quite a scientific study since so many factors figure into what players are drafted, to say nothing of how they actually perform. It’s perfect for the local paper.

So let’s take a look at some newspaper databases. I picked the Miami Dolphins, since at least three fairly large newspapers cover that team (the Miami Herald, the Sun-Sentinel and the Palm Beach Post). I found the Sun-Sentinel’s draft database first – it was teased from the home page – and it is pretty straight-forward, with one problem: it wasn’t working when I loaded it (update: it works now.) The SS’s sister paper in Orlando uses the same app. Search is one good way to start an app like this. It lets you look at specific positions and colleges (provided you know what they are), although just having a search form assumes that your visitors know what to search on. In this case, they probably do.

Next up was the Palm Beach Post’s draft app (disclaimer: I worked for the Post from 1995-1997), which is powered by Caspio. My feelings on news organizations using Caspio are fairly well-known, I think, but this database works and has some good elements, particularly the standardization of positions and use of drop-down menus. The display of the results, on the other hand, is, well, lame. And not indexed in Google (the first result for a search of “miami dolphins draft database” is the SS’s app).

Despite the Herald’s football coverage, I didn’t find a Dolphins draft database on the site. Maybe they don’t have one or are building it now. Maybe they have higher priorities. So those are the current options.

In the interests of not being just another talker, here’s another way to do something like this. The first thing, which is something I didn’t give enough respect to previously, is to devote time and energy to the interface. Like, say, the Indianapolis Star did with its Web feature on Peyton Manning. In particular, check out the interface for the database of passes. But let’s say that you, like me, are graphically challenged or that you have limited resources in the graphics area. All you need to do is be better for your team than DraftHistory.com is.

I believe I have, using Django and 92 minutes of development time (much of which was fixing bad data). Presenting the Miami Dolphins Draft Database that costs nothing except server time and staff time (which, did I mention, was 92 minutes?). Instead of using drop-downs (which I could add), particular slices of the data have their own url, such as first round picks or quarterbacks or players from South Carolina. No, my design isn’t going to win awards, but ask yourself if those other examples will, either.

Just for fun, I’m going to do this as a Rails app, too, to show that it’s not hard to find open-source Web framework options no matter what language you prefer. Comments are welcomed.