What is a Clusterburst?

One of the goals of the AMI University Center is developing visualizations that help people understand the business and economic climate in a region. One particularly frustrating issue is how to compare multivariate data across multiple areas against a standard. Numerically we use location quotient to show how a value for one area compares to another. Location quotient is simply the ratio of a given normalized measure in one area versus the same normalized measure in another area.

The Calculations
This simple calculation lets you see the potential importance of an industry (or any value) to a region. There becomes a bit of a problem however when we have to compare, as in our case, over 20 industries simultaneously over geography and time. In our little study of the North central Kansas with 23 clusters and sub-clusters over 19 counties that results in 437 data points for every year, for the entire state of Kansas – 2,415.

We needed to devise a way to show this data on a map that would allow a user to visually compare LQ data across the region. We were (and are) using Google Earth as the geospatial visualization tool, and we owe a small portion of the development of the clusterburst quirkiness of the Google Earth visual overlay capabilities, and a large portion to the openness of the KML file format. We developed first simple spreadsheets that would write the text files in the KML format that would display our data on a user’s computer.

While unknown to the developers originally the Clusterburst chart is a modification of the Nightingale Rose chart. The chart is also similar to a Radar Chart. While the Clusterburst resembles in a manner a simple pie chart the angles of the wedges remain constant and the scale of each wedge grows or shrinks in size to represent the value. We include a line circle that equals an LQ of one.

What is a Cluster
We feel this visualization best shows two issues that are important for regional economic development. It shows the most clearly the values that represent a LQ greater than one and secondly the irregularity shows the comparative concentration or imbalance of a region’s economy.

The two charts represent two different counties in the state of Kansas, one is a diversified county other is dominated by a single industry.

Transforming Data Using Google Earth
We developed a small program that transforms the data in our database into the KML file used by Google Earth. By using Google Earth we were then allowed to add more layers of information. In the video displayed below we have added markers for the local businesses in the cluster. This lets a viewer see if a particular cluster is dominated by one large business or by many smaller businesses. Additionally by using the third dimension we can show the absolute number of employees as a function of elevation.

BTW with just the slightest bit of encouragement and credit we will share the process we used to develop the actual charts.  Just drop a comment.


© 2018 Innovate Kansas. All Rights Reserved.

Website Design and Development by Imagemakers