I’ve been having a fine time exploring the ProPublica Congress API. I started by making something that would build queries for Congressional TwitterSpace, and that was fun. Then I made something that would be ridiculous and search Congressional TwitterSpace by things like Zodiac sign and % of votes missed. That was even more fun.
But it was also frustrating in a couple of ways: first, these tools only searched Twitter. Second, they only searched the members of the current session of Congress. I wanted to make something that would search more Congressional social media space and search for more representatives across different sessions of Congress.
So I did, and it was the most fun yet. The Congressional Social Media Explorer is available at https://researchbuzz.github.io/Congressional-Social-Media-Explorer/ . You’ll need a ProPublica API key to use it; they’re free at https://www.propublica.org/datastore/api/propublica-congress-api .
The Congressional Social Media Explorer takes your search query, the state for which you want to get member information, the Congress session number (it goes back to 112, which covers 2011-2012), and the branch of Congress. With that data, it builds queries for each representative in that state during that Congress.
(Sometimes it takes a few seconds to show anything since it goes through a pretty big list; don’t expect to hit the button and instantly get the queries. If you’re the impatient type, try searching for Senate members first. Smaller group.)
From these lists of queries you can view a number of things, including Google’s Knowledge Panel for that representative:
You can see what statements matching your query terms have been made on Facebook (and every Facebook search result I’ve seen has been cached by Google, so you don’t need to have a Facebook account to see the entire page. Just look at the cache.)
(Additional queries find posted Facebook photos and Facebook videos.)
Another query takes you to the representative’s Twitter posts:
And finally, you can take a look at the representative’s most recent YouTube videos (these are not searched by keyword, I’m still trying to figure that one out.)
Making this indulged my interest in creating contextual Web searches across discrete slices of time not delineated by the search data (in this case, Congressional sessions.) That kind of search methodology is not easy to casually set up and implement; I suspect I’ll be looking for other ways to experiment with it.
Thanks for reading.