This chapter continued the discussion on mining data from Twitter. After focusing on text and frequencies in Chapter 2, #MiningTwitter - Hashtags, Topics, and Time Series, this chapter focused on the analysis of user connections and interactions. We discussed how to extract information about explicit connections (that is, followers and friends) and how to compare influence and engagement between users.
The discussion on user communities has led to the introduction of unsupervised learning approaches for group users according to their profile description, using clustering algorithms.
We have applied network analysis techniques on data related to a live event, in order to mine conversations from a stream of tweets, understanding how to identify the tweets with the highest number of replies and how to determine the longest conversation.
Finally, we have also shown how to understand the geographic distribution of tweets by plotting the tweets onto a map. By using the Python library Folium...