A couple of weeks ago I had to chance to talk to Jon Udell about one of the projects we've done with 3scale tools to create new APIs -putting some of the great data available at UNDATA online as an API. You can find the API at http://www.undata-api.org/ the conversation on line here.
Since we're an infrastructure I guess it's fair to ask why we put an API like this together in the first place. While without doubt there's the positive benefit of showcasing what our tools can do (the whole site was put together in a couple of days using 3scale's Pro tools and Google's Java AppEngine), there's also a wider purpose in trying to show how easy and valuable it was to get this kind of data on line.
The UN's original site is great for human navigation - you can explore the data sets online and browse at will + search is well organised. Unfortunately though the data can't be downloaded except in zipped bundles of XML which are timestamped at download time - making it hard to grab the bits you really want in real time for any application. The new API certainly isn't perfect - we're still on just a few data sets and the query / return formats are clunky - but this will improve over time and it already allows an application to browse through the data in small chunks. We'll be adding more formats to make that easier and working it into some mashups to show what can be done.
In the long run, the plan is to keep improving the API and open it up for community support - source code for the app will be out shortly + some do's and don'ts for the Java App Engine. We also see it as a growing best practice model for other high value data which should really be out there in API form - and it allows us to make use of latest power tools and get them into people hands for free.
Comments