How to Create a Word Cloud in Tableau

Tableau is a powerful visual data analytics platform. But when it comes to text analysis specifically, it looks like Tableau does not have preset tools and options.

That is only partially true, as Tableau can still create some of the most essential visualizations for text analysis. And one of those is word clouds, which, thanks to Tableau’s top-notch interactivity, can be tailored and customized to be engaging and informative pieces of analysis.

What Is a Word Cloud?

A word cloud is a visualization that shows the frequency of words within a body of text. The frequency of each word is represented by its size relative to the other words. The more frequent a word is, the bigger it is, and that may indicate higher importance, depending on the context.

Word clouds are especially used to analyze customer text feedback, to give snapshots of what customers frequently mention, whether it is good or bad. Similarly, word clouds can be used to analyze social media text content, and open survey responses. There are a lot of possibilities.

Steps to Create a Word Cloud in Tableau 

Let’s work through the steps in order. We will use Women’s E-Commerce Clothing Reviews dataset available on Kaggle. The dataset has a Review Text column, where the clients’ reviews are kept. This is the text column that we will analyze.

Prepare the data

In order to be able to create an insightful word cloud in Tableau with this dataset, we need to perform the following pre-processing steps:

1. Separate each word in the text column into a separate row (this step is also known as tokenizing)
2. Trim and clean the column to take out extra spaces before and after, as well as all punctuation marks and numbers.
3. Convert all words to lowercase to avoid case sensitivity issues.
4. Take out all stop words. There is no fixed list of stop words, but they include pronouns, conjunctions, prepositions, and articles, as well as any words that will turn out to be insignificant and meaningless in our text analysis.

Since Tableau does not have a built-in tool for text analysis pre-processing, we can do this part outside of Tableau, using a tool of your choice, like Python, R, or even Excel’s Power Query.

For example, to preprocess the dataset with R, you can follow the first chapter of our Introduction to Text Analysis in R course.

If you want to cut to the chase, and jump right into the scope of this tutorial, you can use this ready dataset, on which the previous pre-processing steps were already applied [link to download the ready dataset].

Import data into Tableau

Now, let’s connect this dataset into Tableau. Click on the Microsoft Excel connector on the Tableau welcome page and choose the Excel file.

Importing data into Tableau. GIF by Author.

Create the Word Cloud

In Sheet 1, we can create a word cloud through dragging the Word field to the Text mark. This will show all the words (tokens) in our dataset in the same size.

Now we need to differentiate each word according to its frequency of occurrence. To do so, we will:

1. Drag the Word field again, but this time to the Size mark.
2. Right-click on the Word field on the Size mark. From the dropdown list, go to Measure, and then Count. This ensures the word size reflects its frequency.

Tableau will now display a treemap instead of a word cloud. This is not what we want. To fix this, change the Mark type to be Text instead of Automatic.

Now we can see that the words are now a bit differentiated by size, but still not clear. In order to turn this into a proper word cloud, we need to lmit the number of words to the top 15 most frequent ones:

1. Drag the Word field into the Filters.
2. In the Filter window, go to Top, and check By field.
3. Leave everything as is, but change the default 10 to 15.
4. Click OK.

Creating word cloud. GIF by Author.

Congratulations! You have now created your first word cloud.

Customizing the Word Cloud

Now, you might think that the word cloud that we generated is not very informative, and you would be right. To make our word cloud more insightful, we can do some formatting.

Adjusting colors

When all words are black, the word cloud can appear uninformative/unengaging. So, let’s add some colors to it.

Word clouds often use color to visually distinguish each word. To add color, we simply drag the Word field to the Color Mark.

We can then control the color palette of our words, or create our own by choosing a specific color for each word.

Changing colors of word cloud. GIF by Author.

This might be visually engaging, but it is still not very informative. Instead, we can use color to add some meaning to our word cloud.

For instance, we can color the words according to the average rating associated with every word. To do so, instead of having Word in Color Mark, we will use Rating. Then, right-click on the Rating Color Mark and change it to average.

Again, we can adjust the color scale by clicking on the Color Mark.

Coloring word cloud according to a measure. GIF by Author.

Changing font

We can also change the font of our word cloud. To do so, click on the Text Mark and the … button next to Text. Shade the <Word> in the text window, and choose the font type that you want. Click OK.

Adjusting font in word cloud. GIF by Author.

Adding filters

To add more context and engagement controls to the word cloud, we can filter the words according to a specific dimension. For instance, we can show the top words that are mentioned in the reviews of clothing pieces from a specific department. To do so:

1. Drag the Department Name field to the filters.
2. Choose All for now.
3. Click OK.
4. On the Department Name field in the Filters box, right-click and choose Show Filter.
5. On the filter control on the right side, click on the small downward arrow and choose Single Value (Dropdown).
6. Shuffle between the departments to see the difference in the top-mentioned words.

Adding filter to word cloud. GIF by Author.

Advanced Customizations to Further Enhance the Word Cloud

Adding parameters

We built our word cloud to show the top 15 words. But instead of this, we can give the viewer the ability to change the number of the top words to show, within a range that we specify. We can do that through a parameter:

1. Right-click on the Word filter in the Filters box.
2. Choose Edit Filter….
3. In the Filter window, go to the Top tab.
4. Instead of the “15” we specified earlier, click on the dropdown menu and choose Create a New Parameter….
5. In the Parameter window, set a name for the parameter. Let it be Top N Words.
6. At the bottom of the window, we can change the allowed values range. Let’s set the minimum at 15, and the maximum at 100. Leave everything else as is.
7. Click OK in the Parameter window, and OK again in the Filter window.
8. Now, go to the newly-created parameter in the left-side pane, right-click, and choose Show Parameter.
9. Play with the parameter control on the right side and see how it affects the view.

Controlling word cloud with parameter. GIF by Author.

Combining with other visuals

As we can see, word clouds show us the most frequent words in a glimpse, which can give us a sense of what reviewers mention in their reviews. However, word clouds become more powerful when combined with the full review text that each word appears in or a better context and a comprehensive text analysis.

For this, we will combine our word cloud with a table that lists the reviews' full text, in addition to the clothing id for which the review was written, and the rating associated with the review.

In a new sheet we will create a new view that has:

1. Clothing ID and Review Text in Rows.
2. Rating as Text in the table.
3. Change the measure of Rating field to Minimum. Maximum also would do.
4. Let’s call this sheet Review Text.

Creating a table for review text. GIF by Author.

Now, let’s combine the Word Cloud and Review Text sheets in a dashboard. We will need to: 

1. Set a proper dashboard layout.
2. Apply the Department Name filter to the Review Text in addition to the Word Cloud.
3. Set a filtering action from the Word Cloud on the Review Text, so that when we select a word from the word cloud, the Review Text gets filtered to only reviews mentioning this word.

If you are new to building Tableau dashboards, you can follow along this complete tutorial on how to build interactive dashboards in Tableau.

The resulting simple dashboard may look like this:

Word cloud in dashboard. Image by Author.

Word Cloud Best Practices

Choosing the right data

As we saw, word clouds are used for text analysis. Therefore, it makes sense only if we have text pieces, like reviews, that we want to analyze.

In order to create an informative word cloud, the data has to be pre-processed through:

1. Tokenizing the text pieces (separating each word into its own row)
2. Cleaning and trimming white spaces
3. Unifying letter case
4. Taking out stop words

It is also recommended to keep the original text in the pre-processed dataset, as it can add insightful context to the analysis.

Avoiding common pitfalls

The common pitfalls in creating word clouds include:

1. Including all or too many words: Word clouds should only show the top number of most frequently mentioned words. The choice of number of words should also take the available space into consideration, as well as the user’s interest.
2. Using too many colors: As discussed, it is common in word clouds to give each word a different color. However, this can result in a distracting view. Like any other visual, it is better to keep the coloring sparing and meaningful.
3. Ignoring context: Word clouds can be informative and engaging, but they usually need to be combined with filters and other visuals to provide better context and a more comprehensive view.

Optimizing for different devices

Word clouds can be tricky when viewed on different screen sizes. For this, it is always recommended to create separate views for each device type when creating a dashboard with a word cloud, to see and control how it would look like, and to make sure it would not end up too crowded and would still be readable.

For example, the layout of the same dashboard that we created can be customized for phone to look like this:

Mobile view for dashboard with word cloud. Image by Author.

Resources and Further Learning

If you want to explore more creative and fun visualizations in Tableau, keep following our tutorials, like this one on how to create a waterfall chart on Tableau.

Meanwhile, if you want to get serious with Tableau, check out the Data Analyst in Tableau career track. Another great option is our Tableau Fundamentals skill track, which will prepare you for the Tableau Desktop Specialist certification.