Here’s a chart that a colleague made and used many times to illustrate the relationship between rainfall anomalies (an anomaly is the difference from the average) and occurrence of peatland fires in Central Kalimantan, Indonesia.

My main issue with the chart is the confusing double axis and the overuse of loud colors. The eye is constantly forced to consult the legend to figure out what is being shown. The point we want to make with this graphic is: drier-than-normal years have an influence on fire activity. With this in mind, I made this version:

We’ve helped this chart out heaps just by teasing apart the two data sets onto separate axes. Alternating vertical bands help the eye follow the data across different years. And to really hit the point home, I added the circles and dark text.

I also tinkered on the wonderful Many Eyes site to show the data in a slightly different configuration:

This scatter plot shows the hotspot densities as circles of different sizes, which enables one to see immediately that most of the larger circles are tied to negative (drier) rainfall anomalies. If I could customize the chart further, I’d fiddle with the y-axis and change the awful intervals (where’s the zero line?)

What do you think?

©2012 autogeno.us. All Rights Reserved.

• How to lie with charts (0)

Here’s the full picture:

Here’s the picture skeptics and obfuscators want you to see:

The charts show the temperature data from the Hadley Centre, lower-troposphere temperatures estimated by satellites, and the Mauna Loa CO2 data. One chart starts the series back in the mid-1800s. The other picks the last decade.

©2012 autogeno.us. All Rights Reserved.

• WSJ discusses IRI work (0)
• WSJ and COP13 (0)