Text Data Visualization

Data storytellers work with textual data frequently. Let's take a look at how they handle this type of data.

Where do we see textual data?

There are several different text data sources, including web scraping and audio and video data, and may real-world applications for text data (e.g., advertising, customer service, spam filtering, social media analysis).

For example, consider a survey data use case, where individuals participating in a survey can provide text comments corresponding to their opinion regarding the topics of the survey (e.g., employment, mental health, feedback on a product). This data might be leveraged by a Human Resources department to understand employees' perspectives on the current sate of an organization, and inform decisions around new programs/initiatives to launch for employee retention.

Data storytellers use text data mining and visualization in a variety of ways including:

• Pinpointing important trends.

• Summarizing large passages of text.

Visualize text

There are several different ways to visualize text:

• Word clouds

• Word frequency bar plots

• Timeline plots

• Network plots

The first two types of plots are fundamental. We'll be covering their implementation in this lesson.

Word clouds

Word clouds are a popular type of data visualization that involves juxtaposing different words into certain shapes, such as a cloud. The size of the word in the cloud indicates its frequency in the dataset.

Data storytellers typically use word clouds to summarize keywords used in text data to help the audience better understand which words were the most frequent and start to make connections between them. Let's look at a sample implemented using the wordcloud package below.

We have a sample dataset of movie phrases, and here, we use the Text column to generate the word cloud.