Three APIs to measure happiness/emotion on social media:
Addition 11/11/15: Emotion Recognition by Microsoft Project Oxford
We Feel Fine was one of the first social media feelings APIs as far as I can tell, and the teams behind both Hedometer and We Feel say they were inspired by it. We Feel Fine seems great, particularly since it was conceived as a tool for digital art (ideal for the current project I’m working on). That said, it doesn’t really have the time-based granularity I’m looking for, and I don’t really want to just stick with datasets that are dependent on sentences from blog content starting with “I feel”.
Hedometer is attractive since they’re actively adding non-English languages. Their reasoning behind using Twitter as a data source despite it’s non-representative nature is good as well (see About page). Twitter’s not perfect, but it’s a hell of a source if you’re looking for a snapshot of what a subset of people are saying at any given moment.
We Feel seems to give access to the most granular time-based data. I’m looking for hourly samples for a current project, so I’ll probably go with that for now. Their visual explorer on the website is pretty cool as well.