Advanced Visualisations for Text Data Analysis

Explore n-gram word cloud, chord diagram, and a bubble chart, and their implementation in Python

Image 1. Packed bubble chart, Image by author

This article will survey a couple of slightly more advanced graphics for text data visualization in Python. More precisely:

• N-gram word cloud: to display the frequency of higher-order n-grams
• Chord diagram: to show connections between several entities and their strength
• Packed bubble chart: visually engaging display of word frequencies

To illustrate their applications and python coding, I use the classic IMDb 50K Movie Reviews (data license is here). A subset of data was already
pre-processed and cleaned from numbers, stopwords, and special characters.

#1: N-gram word cloud

Standard word cloud from Python’s wordcloud library displays unigrams (single words such as “cat”, “table”, or “flower”). We will explore a slightly more advanced version of the graph, which plots the frequency of bigrams and trigrams (i.e., two and three consecutive words).