Wednesday, April 18, 2012

4/18/2012 English Counts

With all the talk about data, and statistical ranges, and visualizations, and projections, let us never forget that yes, English counts. This map appeared in USA Today on April 13,2012:


It is a wonderful map. The projection allows for minimal distortion of “the lower 48”. The color range is evocative of dryness and drought. The five colors (plus white) are easily discernible. I don’t know the underlying data distribution, but the map agrees with my prejudices about the distribution of usual weather conditions in the US in the springtime.

My only complaint are with the words used to describe the data ranges. “No drought” is good for the first range, and “Abnormally dry” and “Moderate” are fine for the next two, although I don’t really have much of a sense as to whether “Moderate” is drier than “Abnormally dry”, or not. I must take issue with naming the three driest ranges: “Severe”, “Extreme”, and “Exceptional”. I have no idea (without being told in this graphic) if “Exceptional” is drier than “Severe”, or that “Extreme” is not as bad as “Exceptional”. English counts, and the words that you use to describe your data must communicate the relative level of the variable that you are measuring. Five phrases might be “Best”, “Good”, “Middle”, Bad”, and “Worst”. Another effective method would be to include numbers in the descriptions: “5-year Drought”, “10-year Drought”, “20-year Drought”, “50-year Drought”, and “100-year Drought”. There are many phrases you can use to describe data, but please be careful – I find it a bit distressing if you want me to figure out that “Exceptional” is worse than “Extreme”.

No comments:

Post a Comment