Just another WordPress.com site

Adventures in Heat Mapping: Visualizing Frequency


Sometimes if you have a lot of dots on a map you might not necessarily be getting the most bang for your data buck.  The goal of any visualization is effective communication and contribution to understanding.  Visualizations of frequency or density can benefit from Frequency heat maps.


Here are some airports around metro Chicago.  If the visualization required an airport heatmap, then we’d choose the Frequency method to create it because airports are discrete phenomena (as opposed to continuous phenomena -more on that here).

In this illustration I have reference icons at the locations of the airports.  A frequency heatmap will give me a sense of the density of airports in a place -a general fuzzy notion of relative airport-iness.

This brings up an important…point:  Points are philosophical illusions.  In the digital environment they are just really useful hacks upon which we can hang thematic icons, use as interaction hooks, and serve as a locational reference of things that are really too small to see at a given zoom level.


So sometimes it’s helpful to visualize features as light sources who’s luminosity fades with radial distance, like a light bulb.

The pixels at the center of the light bulb are brightest and they diminish until at the edge they have no brightness.


Now, with a constellation of light bulbs, the influence of the items or events merge together with neighbors to give a sense of combined luminosity -how bright with airports places are.

This set of radials are merged together into a single cloud of brightness.  Each pixel is given a sum value from all the contributing light bulbs.


This combined luminosity score can then be mapped to whatever thematic color/opacity range you like.  Here, we map the brightness to a red-yellow-green ramp with decreasing opacity.


The result is a frequency heat map that illustrates how hot with airports places are.

But these airports all count the same, and in reality some things are more important than others.  Could they be weighted by some value to affect the visualization and get a closer modeling of truth?  Yes!

Tying radial distance and/or intensity to a measured value makes frequency heatmaps much more valuable.  Most any type of heat mapped item/event data benefits from weighting, for example:

Airport – Average Traffic
Crime – Severity
Foreclosure – Home Value
H1N1 – Actually, that’s probably better off as boolean
Wellhead – Capacity
Broadcast Tower – Estimated Range
Retail Location – Revenue

Heat maps are lots of fun and can be valuable visual aids.  Be aware, though, that the frequency heat map is not a scientific treatment of data and the falloff rate of the radials are a general notion.  They are, however, among the best and most effective illustrations of density and with the appropriate data their visual impact can be powerful and revealing.

John Nelson / IDV Solutions / john.nelson@idvsolutions.com


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s