Mike Orren Staff

Travis, it's not a matter of paying for data in these cases, but the much more difficult problem of parsing it into something useful -- and the even more difficult problem of doing so across multiple jurisdictions. In the end, some of their work will be open-sourced and hopefully, we'll be able to partner up down the road...

4 months, 4 weeks ago ( Link to this comment | Suggest removal )

Mike Orren Staff

Yeah-- I get ya' Travis. Different records in this case...

4 months, 4 weeks ago ( Link to this comment | Suggest removal )

Mike Orren Staff

cho, I think I know why they stuck to Dallas zips. Same reason we've not implemented some things like crime reports and restaurant scores. Local government is a Tower of Babel. No two jurisdictions report in the same format. It is extremely difficult to take feeds from, say, Dallas and Mesquite and Arlington and make them compatible in any kind of database that makes sense. I'd almost say it is impossible to do so without cooperation from the government agencies involved, something they are generally suspicious about giving, if not downright obstructionist.

Take the campaign contributions database we used to do -- The only way to do it was to hand enter the bulk of the data, which we got by mail and fax from the gummint.

That's part of why Everyblock's work is so important. Crack that problem and there's a wave of not only available, but accessible public info.

4 months, 4 weeks ago ( Link to this comment | Suggest removal )

Jeremy Dunck Staff

jefmelch, I actually did look into that. At the time, greatschools.net was pretty poor, and I looked into the TEA data along with attendance zones.

Unfortunately, I found many problems with the TEA data-- things like a reported low turnover rate in teachers but also a very low seniority. There were enough statistical fallacies that I couldn't have much faith in the reported numbers. It seems to me that people must not be using the TEA data for anything of consequence-- publishing it as required by law, but little auditing of obvious flaws in the reportage.

That along with the total lack of classroom-level statistics made it very difficult to make the TEA information useful to an individual parent or family.

Also, attendance seems to be determined in different ways for different ISDs-- in some cases, there is no geographic rule at all, but simple round-robin allocation of applicants to schools.

Finally, greatschools.net improved quite a bit before I could spend the time to go from proof-of-concept to production on the project, so all in all, we abandoned the project.

If EveryBlock can pull off useful DFW school information, my hat's off to them.

4 months, 3 weeks ago ( Link to this comment | Suggest removal )