Summary

This chapter focused on how to design an infographic to deliver very rich content. LDA topic models result in a set of topics and every topic has a distribution of words. How should we design such an infographic? When we visualize the LDA results, we first want to know the size of a topic, i.e., the percentage of documents for that topic. Then we want to know the similarities or differences between topics. This can be shown by the distances between topics. Then we want to see the distribution of words. It will be ideal to see the distribution of words in the entire corpus, and then be able to choose a topic to see the distribution of words for that topic.

The pyLDAvis library facilitates well-designed interactive infographics. It lets us show the similarities and differences between topics. It shows the distribution of words in the entire corpus, then it lets you choose a topic to see the distribution of words for the topic.

What are other ways to conduct topic modeling...