The purpose of this article is to report a study aimed at systematic assessment of the functionality of commercially available software for the automatic generation of area cartograms. The issue of the wide choice of algorithms developed over the years for generating various types of surfaces has also been raised.
Cartograms (called also anamorphic maps) are constructed by changing the surface area of each spatial unit in step with the corresponding value of the mapped thematic variable (area cartogram) or changing accordingly the distance between the preselected focal point and other points on the map (distance cartogram). Depending on the shape of mapping units, the following three types of area cartograms can be distinguished:
proportional symbol cartograms were the original shapes of mapping units have been replaced with simple geometric shapes such as squares, rectangles or circles (for example, Dorling Circle Cartograms, square cartogram);
continuous regular cartograms where the shapes of areal units on the map resemble the actual shapes of the mapped units but their boundaries have been geometrized to consist of perpendicular sections;
continuous irregular cartograms where the shapes of areal units on the map resemble the actual shapes of the mapped units but unit boundaries are not straight lines.
Cartograms can also be constructed to preserve spatial contiguity (continuous cartograms) or not (noncontinuous cartograms), with the latter often preserving the underlying neighborhood relationships to some degree.
While constructing area cartogram, one needs to decide first the type of a surface to be developed, and therefore, the algorithm to be applied. The next step in map construction is the choice of software, where tool selection depends often on the predetermined type of the cartogram.
The study surveyed five programs for the construction of area cartograms – all available free of charge. Each of those software tools was used to generate area cartograms portraying data from the 2010 presidential election in Poland.
Two groups of area cartograms where generated for the purpose of this study: maps of the entire Poland by voivodships, showing the number of valid votes cast for the two presidential candidates in each voivodship, and maps of the Mazowieckie voivodship by county, portraying the number of valid votes cast for the individual candidates in each county.
The subsequent in-depth assessment of surveyed programs took into account eleven criteria including the number of cartogram types that might be developed using each program, availability of tools for the proper legend construction and display, possibility of supplementing the cartogram with complementary choropleth maps, the option for inserting map labels, the type of spatial data that can be used in the software (reference to points, lines, polygons) and so on.
The study has demonstrated that the tool included in the ArcGIS (Cartogram Utility for ArcGIS) best met the survey criteria scoring 9 points. The application Scape Toad placed second (7.5 pts.), while MapViewer 7 came third (6.25 pts.).
When generating cartograms in the available GIS programs, one should also pay attention to the visual qualities of the generated maps, and in particular, to the resemblance of shapes of spatial units on the map to the their actual geographic boundaries. Since the shape outlines obtained on the map vary depending on the underlying geodetic reference system, the best coordinate system for the mapped area should be selected. However, if such system cannot be used within a given cartogram generating tool, then the obtained cartogram should be exported and refined with some general software package for graphic editing.