This came out of a recent project at Citizen Engagement Lab, where I’m a designer and web developer. We had a request to try and mobilize a group of people’s base of Twitter users to tweet a message at their State Attorney General as part of a campaign. (The project is still in under development, so I leave it at that.)
So how do you get that data, other than painstakingly searching each one?
Well, it turns out you still need to hunt and peck. I took care of that part, but my partner Trevor Cords got it rolling by spending a night scraping most of the data first.
I finished it by combing through and adding the missing ones, which were a lot. And often, it was a judgment call. So I created my own criteria:
- Twitter handle for the A.G. public identity is included where one could be found
- If their Twitter handle seemed inactive, or belonged to an old campaign, and an account belonging to the state office of the A.G was available (@AGmyState), that was included instead
- If the A.G. appeared to have two accounts, one personal (@JoeSmithRadDad) and one pertaining to their role as A.G. (@JoeSmithCalAG), the latter was used
- If the A.G. had two active accounts, both professional, the most recently updated one was used
You can find the data hosted here on Github.