open-austin / project-ideas

:bulb: A place to collect ideas for Open Austin projects
183 stars 25 forks source link

Supreme Court nominee citation exploration #82

Open mscarey opened 7 years ago

mscarey commented 7 years ago

What problem are we trying to solve?

We could make a guide to the most influential passages in opinions written by the judges on the lists of potential Supreme Court nominees that Trump issued during the campaign. I think potential research questions would be (1) what are the legal ideas most associated with each judge? and (2) are any of the judges notably different in ideology from the others? If so, is it possible that the campaign proposed a group of potential justices who appeared homogenous, but the campaign intentionally allowed themselves latitude to nominate somebody significantly different?

Who will benefit (directly and indirectly) from this project?

Voter opinion surveys indicate that millions of people based their presidential votes based on their expectations about the next Supreme Court nominee, which suggests a widespread interest in this topic. It is more of a national issue than an Austin issue, though.

Links to any research/data available/articles

Most of the opinions written by the potential nominees should be available through the CourtListener API. (We wouldn't need every single opinion. We could limit it to opinions with a high "CiteGeist" impact score.)

Because CourtListener has a citation graph, we could also download all newer court opinions that cite one of a nominee's opinions, and all older opinions that are cited by one of a nominee's opinions.

Also, there's this article, which I think reuses ideology scores from other researchers. https://empiricalscotus.com/2016/11/14/trump-court/

What are the next steps (validation, research, coding, design)?

The hardest part might be finding some Natural Language Processing library or tool that can identify similar passages in two separate documents, and getting it to work for our use case.

With the NLP tool, we could find passages in the newer opinions that refer back to similar passages in the potential nominees' own opinions. If those passages aren't also similar to anything that appears in the older opinions cited by the nominees, then they probably contain text that originated with the nominees themselves. So those would be passages most likely to represent the potential nominees' own distinctive ideas.

Even if we manage to create a proper dataset of notable quotations, the project will still just be in the data exploration stage, so we'd have to make a plan for design, and identify a platform to share the research with an audience.

What help is needed at this time?

data exploration, NLP, design, user research

mscarey commented 5 years ago

I'm sure I'll solve this one someday soon