i am working on some geographic analysis of this data. things are going swimmingly so far. i have some fairly interesting analysis of the DIA/DIAA datasets, like # of recruits per state, # of recruits per metro area (MSA and CSA), # of recruits within 180 miles of each school, within 360 miles of each school... et cetera.
the goal is to build a model that represents the best places to recruit. I have some ideas about how to accomplish this.
sooner or later i am going to apply a formula to rank the recruits by talent, and then apply a geographic layer that I call the "Competition Index." The competition index would water down the 'value' of a recruit in Indiana (where competition is high) compared to a recruit in Oregon (where it is low).
the competition index needs a mathematical formula to represent the difference in recruiting power between an elite, a BCS, and a non-BCS, which i don't have yet. I also may incorporate both a 180-mile and a 360-mile analysis, and i'd need to devise a coefficient that would accurately represent the difference between recruiting inside 180, verses 180-360.
what i need, still, is more data. I am only in two worlds, and so i only have two datasets. i don't know quite what sort of sample size i need to even out the seasonal fluctuations by position and geography.
don't ask why i'm doing all this, i'm just a nerd when it comes to maps.
9/22/2010 12:17 PM (edited)