Generating trivia questions from category targeted Wikipedia articles
The basic goal was the to 1. Target Wikipedia articles by category or title, 2. Generate questions using NLP syntatic analysis and pattern analysis to transform an interesting sentence (for our purposes, a sentence that contains a superlative) into a question, 3. Expose a REST API to be able to develop web clients for both computers and mobile. We decided to write our main application in Python We achieved goal #1 by using the MediaWiki REST API since Wikipedia is built on MediaWiki. Goal #2 was achieved by using the NLP library TextBlob to tag parts of speech. Patterns were programmed in based on the patterns we noted over the hackathon. Goal #3 was achieved through the use of Flask to route URLs to our functions, and Angular.js to develop the UX
As a bonus integration, a slackbot was also developed using some of the same DNA in question generation, but instead opting for a wikipedia library that generated the summary rather than entire corpi.