Day two dawned and it was time parse. Fantasy football is a game of information management. Therefore any fantasy football service has to provide an owner with a wealth of information by which they can make sound, season winning decisions.
First up was to update my player database. I had a table of player names and positions left over from some fantasy analysis I did last year. I went through and added the 2012 draft class. I also found lists of the top defensive players and imported those as well. While IDP (or Individual Defensive Player) leagues aren't as common as their offensive (and original) counterparts, they are a growing game type.
Next, I needed to diversify the data for each player. This information is available on the Internet already: things like the latest player news, current mock average draft position, and currently happening trades (since its the offseason these trades are mostly happening among dynasty league owners). But it is fragmented and (unfortunately) buried. A set of autonomous scripts, a family of code that I've come to affectionately call ‘The Pulls', was required.
To just serve these feeds up un-adored, however, isn't enough. For data to become knowledge it is necessary to make logical associations. I need to programmatically match news stories to the players mentioned in them. Here I did a quick shout-out on StackOverflow for a sanity check on an approach.
With ‘The Pulls' now busy at work building context I turned my attention to the logo. "Trade" is an ambiguous concept. A quick peek at a Google image search result shows that plenty of designers opt for representing it with busy, swooshing lines; usually arrows.
Unfortunately I don't seem to have much new to add here. In fact, the initial concept I came up with is not much better than a cliched concept I'd expect from the sites I panned yesterday. However, whenever I work on a design there comes a point where I can't force things. Beyond that point things begin looking worse than when I began. A logo is something that I have to chip away at when I have a fresh eye. As this isn't really needed for a minimally viable product (I'll just put ‘TradeGrade.It' on a title bar) I'm tabling it for now.
Finally, on the onset of the project I resolved to make Git my version control solution. It's because (1) I wanted to learn it in more than just a theoretical way and (2) Github plays wells with a new service called TenXer. The term TenXer is silicon valley slang that has emerged from the idea that great employees are ten times (or 10x) more valuable than average employees. The site TenXer attempts to help developers track their key metrics to help them work better. Since their beta site doesn't yet have the hooks to read from the subversion repositories a majority of my work is in, if I want to see charts up and to the right, I have to use Git.
Tomorrow - back to the interface.