In The Economist e-newsletter 'Off The Charts', about data and visualization, data journalist Elizabeth Lees, writes about mapping and the use of QGIS, the open source mapping software.
How we make maps at The Economist
Since 2016, The Economist has used a geographic information system (GIS) to create most of the maps it publishes in print and online. GIS is a mapping tool that allows us to visualise and manage spatial data quickly and accurately. That is crucial for articles that may have a quick turnaround, such as our Daily charts or Explainers. GIS also allows us to discover and analyse geographical patterns and relationships. To create the dozen or so maps published each week, we source GIS maps and data, which we then load into QGIS—our preferred mapping software.
The QGIS interface with several layers loaded in
Starting from the base up
Every map begins life as a base map in QGIS. A base map is a reference map that provides geographical context and boundaries in which data can be located. Most GIS software offers a range of base maps that are ready to use, but they can also be imported from external sources. For example, we often use Natural Earth, an open-source map database that has accurate maps at various scales. Map layers are typically vector-based shapefiles. These files store the location, shape and any attribute information of geographical features.
When sourcing shapefiles, it is important to make sure all parts of the file are downloaded for it to display correctly. For a world map like the recent Daily chart on the modelled risk of emerging covid-19 variants, we usually download and display polygons (the shapes) of the land and borders of countries (including any disputed borders) to start with. Extra layers, such as transport networks, mountain ranges or lakes and rivers may also be included depending on the context of the story. Shapefiles are also editable, which can be useful when you want to include additional details.
Adding on the layers
To turn the map into an infographic for The Economist, we add data layers on top of the base map. Any data added to our maps must have a spatial element such as degrees of latitude and longitude, georeferenced imagery (such as geoTIFF files), postcodes or country codes, so that it can connect with the map. Additional data can be imported into QGIS depending on the file type (see below). Point co-ordinate data are added as a delimited text layer with a set geometry definition, TIFF files are added as a raster (or grid) layer, or if the data shares a common field with the base map then we use the join tool to tell QGIS where to assign the data.
Additional data layers can be added in QGIS
Once the map and data are all set up, we may decide to do some further analysis to discover and better understand any relationships in the data. This can be done through a number of different GIS tools. QGIS also has an open source plugin portal for user-generated tools that expand the software’s capabilities. Most often the map needs only to show the spatial patterns of the data (such as the choropleth map below) so we skip any analysis and go onto thinking about the visual variables and hierarchy of each map.
A published choropleth map
Beyond the data
Once the data are displayed on the map, we can think about the design. What map projection (see our Off the Charts newsletter from January 25th) and scale work best? What colours should we use? Do we need additional labels? How do we best position the many map elements within our layout? Is the map accessible to visually impaired readers? Although all these aspects can be tweaked within QGIS, at The Economist we usually import the map into Adobe Illustrator for any final touches. Before the finished map is published we share it within our team, as a second (or third) pair of eyes always improves the final result.