ABSTRACT
This chapter employs different strategies to analyze textual data from Twitter. It covers text analysis exploration, and deals with Wordfish (a technique to position texts along an axis). The chapter covers structural topic models, which helps us discover underlying themes in text data. It discusses two recent natural language processing techniques used in political science for unsupervised text mining. In general, using a hashtag indicates interest in a topic, independently of one being in favor or against it. A quick and intuitive way of representing word frequencies are wordclouds. These graphical representations allow to place at the center and with large letters the cases that have greater frequencies. Certain hashtags may increase or decrease in its use through time, depending on the political context. Hashtags can tell a lot about a political debate. Topic modeling is a computational method for automatically identifying relevant word groupings in large volumes of texts.
