Insights

How to Apply the Word Cloud Visual in Power BI

In this article we will investigate what the Word Cloud visual is, how we can create it in Power BI and how to apply it as part of our solution design.

What is a Word Cloud?
Word Cloud is a collection of words with varied sizes and a visual representation of each word’s relative importance in a text. This visual is very powerful for text analysis and will give you a high-level understanding of how often a word has occurred in a text i.e. the size is proportional to the frequency of each word.

How to import Word Cloud in Power BI desktop?
Word Cloud is a custom visual, meaning that it is not part of the built-in visuals, in Power BI. In order to import Word Cloud you need to click on Get more visuals.

Then you can type Word Cloud in the search box and click on the visual

Now you have this visual in Power BI desktop

How to use Word Cloud in text analysis? What type of dataset it is useful for?

When you have a dataset containing reviews of a product, hotel services, etc applying Word Cloud as part of your visualisation is an option. In the below example I show you how to draw an image by Word Cloud using a dataset from Kaggle regarding Hotel Reviews.

When I look at the dataset there is a column named “reviews.title”. So those who stayed in the hotel gave their feedback about their experience. That column can be used to create Word Cloud. By dragging and dropping reviews.title to the Category section the below image will appear (I renamed the column to Reviews)

What you can see in the first glance is that the words excellent and hotel appear bigger as they have been used more frequently in the reviews but the word disappointed is smaller because it was less frequent. From here there are a number of optimisations that can be done:

Removing words that do not convey a meaning:

If you look at the visual there are words such as: the, and, To, not etc. We can exclude these words to make the visual more meaningful. In the formatting section you can make Default Stop Words button on:

It cleanses the visual for you a bit by removing default stop words. Also, if there are words that are still not meaningful you can type them in the Words box below manually. After applying those modifications, the visual looks like this:

You can continue this process until you have more relevant words.

Customising the max number of words:

You might have realized that Word Cloud looks too cluttered. By default, the maximum number of words in the visual is 200. You can reduce this number to focus on a cleaner visual that quickly gives you an idea.

By changing the max number of words to 100 you will see the below image:

Editing the rotate text

Under “Rotate Text” there are boxes that you can set: Max Angle, Min Angle and Max Number of Orientations. By changing the default setting to 0, 90 and 2 respectively Word Clouds looks like the below example.

Mixing Word Cloud with a Slicer:

You can take your analysis to the next well by mixing a slicer and Word Cloud. Let’s go through it with an example. In this dataset there is another column called reviews.rating with a scale of 1 to 5. Based on the scale, 1 means the lowest score (satisfaction) and 5 means the highest. Suppose that we would like to figure out: what the more frequent words are when a customer is highly satisfied or dissatisfied. We can add a slicer based on reviews.ratings and see the outcome.

It is evident that Worst has the occurrence of 49 and looks big in size. So, this is the more frequent word when customers are unhappy with services. Also, you can see words such as Horrible, Terrible, never and so on. Now let’s click on 5:

Based on the above image, when customers give the highest rating Great has the first place in repetition, so it appears as the biggest word in the visual.

Mixing Word Cloud with a Table and Slicer:

In below example I used a slicer based on Hotel Name and added a table containing three columns: Hotel Name, Reviews and Review.Rating. By choosing Greenwich Inn from slicer, it is obvious that word Good has more occurrence in reviews. Also, you can rank ratings by descending order and see for each review what the ranking is.

Conclusion

Word Cloud is a custom visual in Power BI that enables us to interactively analyze more frequent words. A good point about Word Cloud is that it is easily understandable by a non-technical audience and gives you instant insights about the most important words within the dataset.

Have a look at our Power BI trainings, or get in contact with one of our Power BI experts with any of your questions. 

Picture of Natalie Far

Natalie Far

Power BI consultant at Agile Analytics.

Contact Us

Receive a FREE 30-minute consultation with one of our BI Consultants

Find Out