If you can imagine it, you can make it. Imagining is the hard part. Politics, huh? How about a web app that scrapes articles from RSS feeds of political news sites, and makes a word cloud of the most common words in political news. Or, an app that scrapes associations, and draws them with lines whose thickness represents frequency. For example, if the same article has "Jeb Bush" and "Clinton," that's an association. If another article has "Clinton" and "Elizabeth Warren," that's another assocation. So draw three circles connected by lines. If there are twice as many articles with "Clinton" and "Jeb Bush" as there are "Clinton" and "Elizabeth Warren" then the line between their circles is twice as thick. (This is a type of Social Graph. You don't even have to draw it, just scrape the data from RSS feed links and make the database. Tons of stuff you can do with the database.)
I guess my problem then is really that I don't know how data is used. Thanks for the insight!