For a little more than a decade now, DonorsChoose has been linking interested donors with needy public-school classrooms across the United States. It works like Kickstarter: Teachers describe a project they’d like to try with their students, and hope for the best. Projects range from innovative pedagogical experiments that call for expenses outside usual school budgets, to ideas that sound like quasi-random wishlists, to America-damning plaintive requests for basic school supplies in especially impoverished districts. Donors can search by location, by subject, by the poverty of the school’s children (calculated by free/reduced lunch demand), and other factors.
DonorsChoose is in the news this month for launching, with Factual, a new contest: Hacking Education. The Hacking Education contest opens the DonorsChoose API and datasets to developers interested in analyzing these data sets or in developing new apps that would either facilitate donations or explore the data. Accessing the API gives you information about projects currently seeking funding, as well as the functionality of the website. The available datasets include all projects, all donations, purchases of gift cards, the full-text of teacher requests, to website searchlogs, and more. (Obviously, this data has been scrubbed in various ways for privacy.) There’s also a page of complementary data sets, ranging from location-aware services to US education data, to (my personal favorite) Los Angeles School Crime Data.
As Kathi Inman Berens has pointed out, it’s hard not to be excited by the contest’s language, which calls on users “to make discoveries and build apps that improve education in America. Help to shape your school system’s budget by revealing what teachers really need. Build the first mobile app for hyper-local education philanthropy.” And beyond the advertising language, there’s clearly a wealth of data–300,000 projects, submitted by 165,000 teachers from 43,000 districts–here that might interest academic researchers of many stripes. There are reasons to be cautious about the data and contest, too, of course–many districts don’t have much participation, and so the usefulness of the data in any particular location will vary widely. And, of course, since DonorChoose-funded projects don’t include things like “hire 100+ teachers to replace the ones being laid off this year,” there are limits to the ability to meaningfully intervene in budget debates. (Plus, more generally, one would be happier about the contest if DonorsChoose hadn’t previously partnered with the anti-teacher propaganda piece Waiting for Superman.)
What kinds of data sets would help you “shape your school system’s budget”? What kinds of programming challenges might contribute interestingly to public education? Let us know in comments!
Photo by Flickr user Sue Richards / Creative Commons licensed