Advanced Analytical Theory and Methods: Text Analysis - Data Science and Big Data Analytics - page 399

corresponding to 50 five-star reviews extracted from the data. Note that the shades
of gray are only for the aesthetical purpose. The result suggests that customers are
satisfied with the seller , the brand , and the product , and they recommend
bPhone to their friends and families.
Figure 9.10 Word cloud on five-star reviews
Figure 9.11 shows the word cloud of 70 one-star reviews. The words sim and
button occur frequently enough that it would be advisable to sample the reviews
that contain these terms and determine what is being said about buttons and
SIM cards. Word clouds can reveal useful information beyond the most prominent
terms. For example, the graph in Figure 9.11 oddly contains words like stolen
and Venezuela . As the Data Science team investigates the stories behind these
words, it finds that these words appear in 1-star reviews because there are a few
unauthorized sellers from Venezuela that sell stolen bPhones. ACME can take
further actions from this point. This is an example of how text analysis and even
simple visualizations can help gain insights.
Figure 9.11 Word cloud on one-star reviews
TFIDF can be used to highlight the informative words in the reviews. Figure
9.12 shows a subset of the reviews in which each word with a larger font size
corresponds to a higher TFIDF value. Each review is considered a document. With