Predicting Elections with Twitter
Before you get on your horses:
- I read the article about You Can’t Use Twitter to Predict Election Results.
- I also read the paper from Daniel Gayo-Avello from the University of Oviedo (Spain) in which he states that all past twitter election ‘predictions’ were either not really predictions, or flawed in some way (see p. 2).
In his paper, however, mr. Gayo-Avello also added:
- “Recommendations for Future Research regarding Electoral Predictions using Twitter Data” (p. 3), and
- “Core Lines of Future Research” (p. 4).
So, unless I’m reading this wrong, it might be possible to predict the outcome of election with twitter, given we follow the proper procedures to do it.
So, now we know it might be possible, and since the goal “Predicting Elections with Twitter” is clear to almost everyone, I’d like to take this as the starting point for a Big Data series on this blog.
Disclaimer: I’m not saying that at the end of this series we’ll have a full-blown cristal twitter ball, but at least we will have learned a lot about Big Data and all the challenges that come with it.