PHLT 8076 Week 7 Thematic Map Type Rationale Example

Shaded Tracts, Not Circles or Dots: Arguing for a Choropleth of Late-Stage Percentages and Explaining Why Graduated Symbols and Dot Density Were Set Aside

Student Name

Walden University

PHLT 8076: Fundamentals of Mapping and Geographic Information Systems for Public Health

Instructor Name

Month Day, Year

Shaded Tracts, Not Circles or Dots: Arguing for a Choropleth of Late-Stage Percentages and Explaining Why Graduated Symbols and Dot Density Were Set Aside

The Choice to Be Made

The composite county's late-stage breast cancer data could reach the page as shaded tracts, as circles scaled to counts or as dots, and each would tell readers something different. This paper argues for a choropleth of the percentage of diagnoses found at a late stage, using the classification chosen in Week 4, and explains, map type by map type, why the alternatives were set aside for this purpose. The question the map serves remains the one set in Week 1: in which tracts are diagnoses most often late, and does that pattern follow access to screening?

Option 1: Choropleth of Percentages

A choropleth shades each area by a value, and it works best when the value is a rate or proportion that does not depend on the area's size. The percentage of diagnoses found late is such a value: it compares tracts directly, whatever their population. A choropleth also invites comparison with other tract-level layers, such as travel time or social vulnerability, because they share the same units. Its known weaknesses are two. Large, sparsely populated tracts draw the eye more than small dense ones, and the shade of a tract implies that the value applies uniformly across it (Monmonier, 2018). The first is addressed with an inset for the urban east side; the second is inherent and will be stated in the caption.

Option 2: Graduated Symbols

Circles sized by the number of late-stage diagnoses in each tract would show where cases are, which is the question Week 1 argued the dashboard already answers. Symbols handle counts well because they avoid the area problem: a small urban tract with many cases gets a large circle no matter its size. But the map's question concerns risk, not volume, and circles sized by counts would again make the populous east side look like the problem. Circles sized by percentages would be harder to read than shades and would overlap in dense areas. Slocum et al. (2009) recommend proportional symbols for totals and choropleths for derived values such as rates, a rule that points directly to the choropleth here.

Option 3: Dot Density

A dot-density map places dots within each tract, each representing a set number of cases, to show where cases are distributed. It conveys density well and is visually intuitive. For health data it raises two problems. Dots placed at random within tracts imply locations that are not real, and readers may assume each dot marks a home. With confidential registry data, any method that suggests individual locations invites concern about privacy, even when dots are randomized. And like graduated symbols, dot density shows counts, not risk.

The Comparison Summarized

Table 1

Three Map Types Judged Against the Map's Question

Note. Assessment refers to the composite county map of late-stage breast cancer, 2018-2022.

Combining Types

One combination was considered seriously: a choropleth of percentages with small circles for mammography facility locations. That pairing uses each type for what it does well, shades for a proportion and points for real locations, and it lets readers see the relationship between late-stage percentages and screening access at a glance. It will be used. A choropleth with circles for counts layered on top was rejected, because two quantitative encodings on one map compete for attention and confuse readers who must decode both legends.

What the Choropleth Still Requires

The choropleth's honesty depends on choices made elsewhere in the course: the classification method stated in the legend, hatching for tracts with fewer than ten diagnoses, counts available in a companion table and a caption reminding readers that a tract's shade describes all its women together, not any individual. Elliott and Wartenberg (2004) caution that area-based maps describe groups and cannot be read as statements about individuals, the ecological fallacy, and the caption will make that point plainly.

Testing the Choice With Readers

Before the map is finalized, the choropleth and the graduated-symbol version will be shown to five staff members and two board members who were not involved in making them, with one question: where are women most likely to be diagnosed late? If readers of the symbol map point to the east side and readers of the choropleth point to the south, the test confirms that the map types answer different questions as argued here. If both groups point to the same place, the difference matters less than expected, and the paper will say so. The test also checks whether readers notice the hatching on small tracts and understand it, since a design feature that readers overlook does not protect them from overreading unstable values. If the hatching goes unnoticed, the legend will be revised to explain it in plain words. Results from the test will be summarized in the paper's appendix, along with any changes made in response. Testing with readers who did not build the map is the only way to learn what a map actually communicates, as opposed to what its designer intended, and for a map meant to guide a screening program's outreach, that difference is worth the hour the test requires. It is cheap insurance against a misread map. The results will shape the final legend. A brief test of this kind costs little and guards against the designer's assumptions about what readers will see.

Conclusion

For a question about where late-stage diagnosis is most likely, a choropleth of percentages with facility locations shown as points is the most direct and least misleading choice. The alternatives answer different questions well; they do not answer this one.

References

Elliott, P., & Wartenberg, D. (2004). Spatial epidemiology: Current approaches and future challenges. Environmental Health Perspectives, 112(9), 998-1006. https://doi.org/10.1289/ehp.6735

Monmonier, M. (2018). How to lie with maps (3rd ed.). University of Chicago Press.

Slocum, T. A., McMaster, R. B., Kessler, F. C., & Howard, H. H. (2009). Thematic cartography and geovisualization (3rd ed.). Pearson Prentice Hall.