Create classes to analyze reddit data much like we are doing with twitter data. Instead of fetching data from twitter, I would just read local files generated by your reddit fetcher. You could then extract the commonly usable classes from the twitter.stats package into a new stats package, and implement them in both the twitter.stats package and a new *reddit.stats package. Most code should be usable for both use cases. Coordinate changes with @christian176 and @DavidStei, as they are working on the twitter part.
Since we can only fetch so many tweets a day, we should cache tweets locally. I would implement a configurable file cache (which we can set to our google drive folder), which saves tweets in a file named after the user's twitter handle. Before fetching, we should always check if this file exists and if it doesn't we will create it. I would implement the Serializable interface in the TwitterUser class and then use that to save and restore the instances.