Maps in reports: designing maps that inform

Maps are popular in annual reports, impact reports and policy briefs. A map of project districts looks authoritative and fills a page well. But a map is only useful if it helps the reader understand something they could not see as quickly in a sentence, a table or a bar chart.

The Financial Times Visual Vocabulary, a widely used guide to choosing charts, says spatial charts should be used only when precise locations or geographical patterns in the data are more important to the reader than anything else.¹ Before you commission a map, write down the question it answers. "Where do we work?" is a valid question, but it needs a simple locator map, not a coloured data map. "Which districts have the lowest vaccination coverage?" may need a thematic map. "Did coverage rise this year?" is probably better answered with a line or bar chart.

Most maps in reports fall into three types. A locator map shows where things are: offices, project sites, a refugee camp. A choropleth map colours each area, such as a province or district, according to a value. A proportional symbol map places a circle on each area and sizes it by a value.

The choice depends on your data. The Visual Vocabulary advises that a basic choropleth should always show rates rather than totals and use a sensible base geography, while proportional symbols are suited to totals rather than rates.¹ Datawrapper's guidance makes the same point: choropleth maps work best for relative data, and absolute numbers are better shown on a symbol map.²

In practice, this means the number of people reached in each district belongs on a symbol map, while the percentage of households reached belongs on a choropleth map.

Raw counts mislead on a choropleth map because populous areas will almost always have the highest numbers. A district with many people will look like a hotspot even if its rate is average. Datawrapper notes that raw counts cannot be compared across regions without population figures, which is why rates are more useful.² Convert counts to a percentage or a rate per 1,000 people before you design anything.

There is a second trap. Choropleth maps show how much geographic area is affected, so large, sparsely populated regions can draw the eye more than small, crowded ones.² In Rwanda, for example, the three districts of the City of Kigali are small on the map compared with large rural districts. If the story is about people rather than land, add a table, label key areas directly, or consider a symbol map. Datawrapper suggests population cartograms as a fairer view of where people live, though they work best when readers already know the geography.²

Pick the colour scheme to match your data. Datawrapper recommends a sequential scheme, from light to dark, to highlight high values, and a diverging scheme, with two hues meeting at a neutral midpoint, to highlight both extremes.² It also advises making sure there is a clear difference in lightness across the scale, and using colours that people with colour blindness can tell apart.²

Next, decide how many steps, or classes, the scale will have. Datawrapper's Lisa Charlotte Muth describes this as a compromise between honesty and usefulness.³ More steps add contrast, but they can also make readers think differences are stark where they are not.³ In our own practice, four to six steps suit most report maps.

Spend time on the key. Datawrapper advises that keys for sequential schemes show the colours for the lowest and highest values, and recommends equal value intervals such as 0, 25 and 50.² Round the numbers in the key, and use the same scale on every map you want readers to compare.

Many readers will meet your report as a PDF or web page, some through a screen reader, and some with colour vision deficiency. The W3C explains that its Use of Color success criterion is intended to ensure that all sighted users can access information conveyed by colour differences.⁴ On a map, this means adding direct labels or values for key areas, using patterns or outlines where needed, and never relying on red and green alone.

The W3C lists maps among the complex images that contain more information than a short phrase can convey, and says such images need a two-part text alternative: a short description that identifies the image, and a long description that conveys its essential information.⁵ For a report map, the short description can name the map and its main finding, and the long description can be the underlying data table, placed in an annex or next to the map.

Maps can carry political meaning. Disputed borders, territories and place names are sensitive, and an error can cause real problems for an organisation working with governments and UN agencies. Use base maps from official or trusted sources, and check place names against the spelling your partners use.

UN agencies commonly print a disclaimer on their maps. The map disclaimer used by United Nations Volunteers, for instance, states that the designations employed and the presentation of material are in conformity with United Nations practice, and do not imply the expression of any opinion on the legal status of any country, territory or area, or on the delimitation of its borders.⁶ If you are producing a report for or with a UN partner, ask which disclaimer wording they require before the map is designed.

A good report map works on its own, because readers often skim a report and stop at the graphics. In our own practice, we give each map a title that states the finding, such as "Coverage is lowest in the northern districts", rather than a label like "Coverage by district". Below the map, add a source line, the date of the data and any notes on how rates were calculated.

Remove everything that does not help the reader. Roads, rivers and terrain shading are rarely needed on a thematic map, and heavy borders compete with the data. Check that the map is still legible when printed in black and white and when viewed on a phone, since many partners will open the PDF on a small screen.

• Title that states the main finding.
• Key with rounded, evenly spaced values.
• Direct labels for the areas the text discusses.
• Source, data date and method note.
• Disclaimer where a partner requires one.
• Short and long text alternatives for digital versions.

• The map answers a clear question about place
• Map type matches the data: rates on a choropleth, totals on symbols
• Four to six colour steps with a clear lightness change
• Colour-blind-safe palette and direct labels for key areas
• Title states the finding; key, source and date included
• Short and long text alternatives, with the data table in an annex
• Boundaries, names and disclaimer checked with partners