What is cartography?
Cartography is the practice of designing maps so that people understand them. A GIS can produce a map in seconds. Whether that map answers a question, and whether a reader reaches the right conclusion from it, depends on design choices.
Start with the reader
Before opening a layout, decide who will read the map, what question it answers, and where it will be used: a printed A4 page, a slide, a web page. A map for an engineer who needs coordinates is different from a map for a public meeting.
The single message of the map decides what goes in it. Everything that does not help the message should be removed or made quiet.
Map elements
A finished map usually has:
- a title that says what the map shows
- the map itself, with a clear extent and scale
- a legend that explains only the symbols a reader cannot guess
- a scale bar, since printed or resized maps change scale
- a north arrow, when north is not up or the area is not obvious
- the data source, the date, and the coordinate system
- the author and any licence text that the data requires
An inset or overview map shows where the main map sits in a wider area.
Scale and projection
Scale decides how much detail fits. A large scale, such as 1:10,000, shows a small area with detail. A small scale, such as 1:1,000,000, shows a large area with little detail. Set a clean scale and check the map still reads at the printed size.
The projection affects shapes, areas and distances. Pick one that suits the area and the purpose, and state it on the map.
Showing data honestly
Classification
When a numeric field is shown in classes, the breaks change the picture.
| Method | How it splits the data | Suits |
|---|---|---|
| Equal interval | Same-width ranges | Data spread evenly, and a legend readers can interpret easily |
| Quantile | Same number of features in each class | Comparing rank, but it can hide large gaps |
| Natural breaks | Groups with the largest gaps between them | Uneven data, though class limits are hard to compare between maps |
| Manual | Breaks you set | Meaningful thresholds, such as a legal limit |
Five or fewer classes are easier to read. Try two methods and see whether your message survives.
Counts and rates
A choropleth map colours areas by a value. Colouring areas by a raw count, such as the number of hospital beds, mostly shows which districts are large or populous. Use a rate or density instead, such as beds per 1,000 people.
Colour
Pick the colour ramp from the type of data.
- Sequential ramps, light to dark, suit values that run from low to high.
- Diverging ramps, with two hues meeting at a meaningful midpoint, suit values above and below a reference, such as change from a baseline.
- Qualitative palettes, with distinct hues, suit categories with no order.
Avoid rainbow ramps for ordered data, because they have no natural order and some bands look more intense than others. Check the map for colour-vision deficiency, and make sure text has enough contrast.
Labels and hierarchy
Readers look at the strongest elements first. Make the subject of the map the strongest, and put base layers such as roads and boundaries in light greys. Label only what helps. Place labels so they do not cover the data, and use a halo or buffer around text over busy imagery.
Print layouts and atlases
A print layout is a page with a fixed size where you arrange the map and its elements. It reads from the project, so changes to symbols carry through. Locking the layers in a layout map stops later edits to the project from changing a finished page.
An atlas repeats one layout for each feature in a coverage layer, such as one page per district. A label can show the name of the current feature, and the map frame moves to fit it.
Example: parks and lakes of central Bengaluru
The QGIS layout tutorial builds a page with a title, a map of Sentinel-2 imagery with park and lake polygons, a legend with renamed entries, a scale bar, a north arrow, a coordinate grid, an overview map and a credit line. It then turns the same layout into an atlas with one page per large park.
Common mistakes
- Starting to arrange elements before deciding what the map is for.
- Colouring areas by raw counts.
- Using a rainbow ramp for ordered data.
- Too many classes, or classes picked without checking alternatives.
- Leaving out the scale, the source or the date.
- Letting the base map compete with the subject.
Glossary
| Term | Meaning |
|---|---|
| Choropleth map | A map that colours areas by a value |
| Classification | Grouping a numeric field into ranges |
| Sequential ramp | A light to dark colour scheme for ordered values |
| Diverging ramp | A scheme with two hues around a meaningful midpoint |
| Print layout | A fixed-size page for arranging a map and its elements |
| Atlas | A series of map pages generated from a coverage layer |
Practise it
The tutorial Print layouts and map design builds the page step by step, including the grid, overview map and atlas.
Test yourself
1. Why is a raw count a poor choice for a choropleth map?
Large or populous areas will have large counts, so the map mostly shows size. A rate or density removes that effect.
2. Which kind of colour ramp suits change from a baseline?
A diverging ramp, with the two hues meeting at the baseline.
3. Why include a scale bar on a printed map?
A printed or resized map changes scale, so the number alone can mislead. A bar scales with the page.