Relationships in data are everywhere; from the products we buy to the places we go, relationships in data make sense out of seemingly random data.  Take for example a market basket analysis.  This type of analytics determines that products are likely to sell with other products, for example, peanut butter and jelly.

chord chart

The previous way to visualize this type of data was as 2 separate charts, hopefully, one of which was prompted so that a user could cross compare.  Some more advanced visualizations would let one select affect the other.

Chrod Chart

This methodology has worked for years, but there is a newer visualization to represent this relationship; the chord chart.

The Chord Chart

We’ll start with the neuroscience of the chart.  The charts works on 3 encodings which are position, area, and color hue.  The position of the nominal category around the outer ring shows which items have the least to most interval value.  The width, or area, of the connection between one nominal value and another various depending on the value within the relationship.  The color hue helps designate the nominal value with the greatest overall value in the relationship.

Essentially, the outer ring works like a donut chart; showing which category has the greatest percentage of the whole. The chords show how strong the categories are related.  It mimics the process of a data matrix.  If you were to take Product2 and pivot it, you would end up with the matrix below.

chord chart

If we look at the table, there are too many data points to see a trend or correlation quickly.  Using multiple charts, the comprehension is better, but it requires a user’s input to farther the investigation.  The chord chart provides the graphical representation needed for the data matrix.

Representing data visually is tricky.  We want to convey a message and that message needs to be clear.  We know now that an audience is easily distracted with limited time and attention for your data and know that data can come in all shapes and sizes.  In this blog series, we have identified how to communicate to our end users to overcome their cognitive limitations, both with different data encodings and modern visualizations that rely less on the user’s ability to understand the nuances of the data.

Free eBook Download

In the eBook “Having a Conversation with Data”, learn what the current BI infrastructure has been and associated challenges with the traditional approach. How important the user experience is in order to best maximize data’s value (think visualizations!!) to your organization and how to gain a competitive advantage with modern analytics platforms. Click here to download.