First APIs Available
Already the Sunlight Mash-Up Labs announced in May is striding toward my fantasy of one-click political influence disclosure. Last week, Lab Co-director, Greg Elin, guided me through the results of a week of "hacking" with Mike Krejci, lead programmer for The Institute of Money in State Politics. Supported by a small grant from the Sunlight Foundation, Greg went to Portland, Oregon and helped Mike begin work on The Institute’s "web services API".
The Institute tracks campaign finance data on some 18,000 state-office candidates each election cycle and now manages a database of some 14 million records spanning many years. Even though The Institute makes this data available via its respected FollowTheMoney.org web site — which is pretty amazing when you think about it — the fact is getting at that data can be cumbersome, especially when you are on a different web site. As it is now, looking up information on your state candidate means leaving whatever website you are on and going to FollowTheMoney.org and searching through various pages to look up the data you want.
Web services API changes this picture dramatically. According to Greg, a web services API (short for Application Programmer’s Interface), "is a machine-friendly interface to a web site’s underlying complex database and application." By adding a web services API to their web site, The Institute is making it significantly easier for programmers at other web sites to dynamically incorporate The Institute’s data into their own web-based applications. And that means in the future you and I won’t have to change web sites to see the data that matters. It will already be there.
To give us non-ubergeeks a sense of this future, Mike and Greg mashed-up a few web page "widgets" which remotely search The Institute’s data. You can try one here. You can search by state, year, office, won/lost, party and even candidate without ever leaving the web page or even reloading the web page. Your search is automatically sent to The Institute’s API in the background which delivers the results dynamically into the page at which you are currently looking.
The ability to easily integrate data from one web site into another really changes the big picture. There’s simply too much data for a single entity to manage. It simply takes too long to bounce from site to site to research subtle patterns of influence buying. But allowing summary data, or detailed data, to more easily move between data silos creates the means to browse — and compare — hundreds of million data points simultaneously. Pretty neat.
Greg tells me Mike still has work to do before the APIs are ready for public release, but that Mike made enough progress they are ready for limited trials with The Institute’s partners. I can’t wait to make further announcements.