Bivariate Choropleth Maps
Overview
A bivariate choropleth map is a type of map where the color applied to each areal unit (such as a US county or state) is based on a grouping scheme for a combination of two variables.
The color palette is based on a 3 x 3 grid, where each pixel corresponds to a particular bivarate grouping. For tertiles, there are nine possible bivariate groupings: “LL”, “LM”, “LH”, “ML”, “MM”, “MH”, “HL”, “HM”, and “HH”. The letters “L”, “M”, and “H” correspond to the first (lowest), second (middle), and third (highest) groups based on the tertiles.
Consider the following example for five counties (ordered alphabetically by county name) with the diabetes indicators “Diagnosed Diabetes (Percentage)” and “Obesity (Percentage)”, along with their bivariate tertiles at the National and State geographic levels:
County | State | Diagnosed Diabetes (Percentage) | Obesity (Percentage) | Bivariate Tertiles (National) | Bivariate Tertiles (State) |
---|---|---|---|---|---|
Autauga County | Alabama | 9.5 | 29.6 | HM | LM |
Baldwin County | Alabama | 8.4 | 28.3 | MM | LL |
Barbour County | Alabama | 13.5 | 29.3 | HM | HM |
Bibb County | Alabama | 10.2 | 23.1 | HL | LL |
Blount County | Alabama | 10.5 | 27.8 | HM | LL |
The first entry in the table is Autauga County, Alabama, where the estimated percentage of diagnosed diabetes is 9.5% and the estimated obesity percentage is 29.6%. The National bivariate tertile group “HM” indicates that Autauga County is in the highest (“H”) group for diabetes and the middle (“M”) group for obesity. Likewise, Baldwin County, Alabama (= “MM”), is in the middle (“M”) group for diabetes and the middle (“M”) group for obesity.
Differences in the geographic level can result in very different groupings. For example, relative to other counties in the US, Autauga County has a relatively high percentage of adults with diabetes, but relative to other counties in Alabama, its percentage of adults with diabetes is relatively low.