12.5. Design Activity

Come up with at least two different theoretical sets of rules (recommendation algorithms) for what would make a “good” tweet to recommend.

• Consider all the information you could get about a Tweet, both from the Twitter API, but also information Twitter has internally. For example

  • Tweet engagement

  • Information about the user

  • Information about the topic (that we can try to guess)

  • Other data mining strategies (like Sentiment Analysis)

• Compare your two different strategies, and think about how users might try to behave in order to game the algorithm