Choosing a Visualisation Option

Having uploaded and processed your data to Caplenawhat’s next? Data analysis is most effective when it explains key points through graphs 🚀 

Creating a bar or pie chart seems simple enough, however, the challenge lies in making sense of it. In light of this, it is crucial to understand why you should visualize specific data and how to best translate it with a particular graph.

The Treemap is this month’s focus. Want to know more? Okay, curious one, read on😉 📖

What is a Treemap Chart? 🌳🗺️

Treemaps are used to present large amounts of hierarchically structured data, and, since Caplena’s text analysis results are hierarchical by nature, you can be pretty certain that most of your data will be suited for this type of chart. But – let’s not get ahead of ourselves and start with the basics first ☝️.

A Treemap is a type of graph where quantities are compared and hierarchical structures are visualized spatially through rectangular segments. Within the visualization, these segments vary in size and color and are sorted by a numerical variable. Rectangles contain smaller rectangles to represent the different ‘levels’ ⬜◻️▫️. The main idea with this type of graph is to break up the data so that large and small units can be quickly identified 🤓 🔎.

The concept of a recursive construction allowing hierarchical data extension first originated in the early 1990s by Professor Ben Shneiderman (see example below for one of the first Treemaps).

How is the Square Size Determined?

Hard disk space usage visualized in TreeSize, the software first released in 1996

How the rectangles of Treemaps are sized depends on the algorithm which is used to define their size. There is an order to Caplena’s rectangle sizing, beginning with the largest square in the top left corner and ending with the smallest in the bottom right corner. The size of rectangles is determined by the sum of their contained areas. Another useful feature within Caplena is that you can interact with the visualization as well. Simply click the desired square to see the coded verbatim responses associated with each square, or double click on an area to zoom in on less frequent codes.

Still Unsure? Here’s an example.

Imagine you are looking at a world map 🌍, and that the continents are represented by squares, instead. Let’s take Asia, one big square, which is further segmented into 48 squares to represent its countries. Next, we could take Europe – this overall square would be a lot smaller than Asia. Europe would have 44 respective smaller rectangles within it representing its respective countries. See the example below.

Source: Plotly

Is it not easier to comprehend the continents and their countries using such a visualization? Or would you rather scroll through 48 lines summing up a single continent? In most cases, the Treemap visualization of a continent would be preferred. This is because we do not need to go into great detail about every country to understand the continents puzzle pieces; we simply want to acknowledge their existence as a microcosm.

Treemaps for Text Analysis

As we are hopefully getting on the same page, let’s bring our attention back to Caplena. Caplena is a text analytics platform helping you reach insights from unstructured text. The gold standard in text analysis is to build a topic-hierarchy (aka the “code-frame”) and assign the topics (aka “codes”) to verbatims.

After uploading your free-text responses, Caplena’s AI will search for hierarchies, patterns, trends and gain knowledge from them. Once Caplena is finished, you are left with a set of organized data that has been assigned codes. Thus, the whole idea behind Treemaps is very well suited to text analytics data 🙌.

Treemaps in Caplena

What should you use Treemaps for?

Having established that Treemaps are good for text analysis, let’s look at Treemap do-s & don’t-s in a more general context.

Avoid 👎

Go Ahead 👍

It is not recommended to use Treemaps for data with a wide range of magnitude, values with negative ranges, or for gauging values based on the length of the plot. Generally, avoid Treemaps for segmentation, as for example, comparing distribution by responded age.

It is recommended to use Treemaps when the data is hierarchical (which entails pretty much all of Caplena’s data). Treemaps offer the main benefit of allowing one to identify the most important elements in an organizational structure at a glance.

We could go on and on – but it’s always best to learn by doing 💪, so we highly suggest you Book a Demo or Start your Trial here (no credit card information necessary!) Try out the different data visualization options, including the Treemap, and let us know how it goes. We are always happy to assist you with any questions you may have 😊.

Enjoyed this short read? Let us know what you’d be interested in next!

Just email our Head of Marketing sheila [at] caplena.com

Previous ArticleNext Article