Detailing on choropleth maps
In epidemiology and public health we often need to represent health problems using rates rather than individuals. Clearly when mapping rates (ill/population) we can’t use spot maps (e.g. John Snow’s cholera map)but need instead to resort to choropleth (area) maps. Examples cancer maps are in ET’s first book. The problem with choropleth maps is that the shading would tend to obscure or obliterate any other geographic features. Accordingly, the value of the map as a layered instument for analysis, comparisons, relationships etc. is lost. It seems to me that it would be possible given a careful choice of colors or shadings to add additional layers to the map showing geographic (rivers, roads, high altitude) or environmental features that might be related to the disease distribution. However, I can find no examples (except for political boundaries or an ocasional superimposed spot) even in cartography texts. Do you know of any example choropleth maps showing other critical geographic features? — any suggestions on shading or coloring?
Thanks,
Bob Fontaine
If this is a question of simply overlaying and/or comparing geographic areas (i.e. polygons) with other geographic features (points, lines, or polygons), it can easily be accomplished with basic Geographic Information System software. Most GIS software will allow the display of polygon fills as solid, patterned, transparent, or semi-transparent. Depending on your source data, you could also use the choropleths as your base and layer other data on top. Links to many excellent GIS packages can be found in another thread.
A more manual/analog layering approach of geographic data for analysis is described in Ian McHarg’s classic book Design With Nature.
I work on detailed chloropleth maps (the geologists here call them isopach maps) all day long. As far as a design doctrine is concerned, I use the color fill to convey the macro view, while the lines, points, and text labels all work only on close inspection. I typically use a saturation scale to convey the magnitude or intensity of a variable, which allow for two or (sometimes) more variables to be mapped in exclusive (non-overlapping) areas. For instance, the thickness of a geologic formation might be shown as a scale of light blue to dark blue within the client’s property boundaries, and then shown as a scale of light yellow to brown outside the property.
For software, I use ArcMap from ESRI, Inc.
I am working with health data as well and my background is in GIS and geography.
It is really hard to put a lot information on thematic maps and try not to confuse the reader. In the link below I tried to put some other geographical features on a map showing diabetes prevalences for Southwark, a Borough in South London.
http://www.spatial-literacy.org/health/map_gallery/slides/Atlas22.html
The dots, at practice level, show the standardised prevalence ratio (observed/expected) for diabetes using the national average as base. So values of the index above 100 means that the practice is above the national average.
The surface underneath is a simple prevalence rate surface (number of cases/population at risk). In this case I had to aggregate data collected at practice level in way that they were geographically consistent.
In the legend is called unadjusted prevalence ratio, but It would be better to call it unadjusted prevalence rate, as ratio as usually rate of rates.
Basically, the map present two measures of the same condition: one comparing the national average at practice level, the other considering the geographical distribution at local level.
As you can see I also display tube stations, major roads and parks to allow a sort of visual dasymmetric mapping.
Please feel free to see the other maps in the gallery.
I hope it can help in this discussion and I am waiting for your comments.
Imago Urbis: Giuseppe Vasi’s Grand Tour of Rome.
This is a map overlay site including the highly detailed 1748 map by Giambattista Nolli.
via Metafilter.