Simple, informative and visually coherent maps can be strong allies in a pandemic: 10 tips to deliver

These guidelines have been jointly written by CartONG & MapAction.

The ongoing COVID-19 crisis has impacted in one way or another the work of nearly all humanitarian actors. With the global, national and local dimensions of the pandemic, static and interactive maps and dashboards are more valuable than ever before in communications and decision-making and a large number have been created. As Fawad Hussain Syed, co-chair of the GIMAC (Global Information Management, Assessment & Analysis Cell on COVID-19), remarked during a meeting with its members: “Everybody has a dashboard now … Every organisation sees it as a visibility opportunity to have a dashboard.”

When so many information sources are available, creating data visualisations that are inadequate or don’t meet a clearly defined need just adds to the noise and may even result in misinterpretation or wrong directions being taken. Sometimes good intentions – such as an attempt to include everyone’s suggestions or meet everyone’s needs in one map – may turn bad, resulting in an overcrowded product that is hard to read. Likewise, a small, innocent mistake, such as forgetting to include the data release date on a map, may also result in wrong conclusions being drawn.

In order to help enhance the quality of maps used for humanitarian response, both during the COVID-19 situation and in the long-term, CartONG and MapAction have put together this quick guide giving some simple and concrete tips on how to deliver effective maps and avoid common pitfalls. Many of the points raised in this document also apply in the context of other types of pandemics, natural disasters and humanitarian emergencies.

You have been asked to create a map showing health service coverage against case load. After talking with colleagues, you have a range of requests for things to be included: one would like to see road networks, another the travel time between distribution centres, and yet another some verbal feedback from community members. The result is a map in which you can barely see the background as it’s covered with dots, polygons, lines and labels. A map that is basically unreadable serves no one.

It is important to keep the map simple. Instead of cramming everything into one map , consider creating separate maps for the logistics and community engagement teams for example. You may also want to consider adding visualisations, such as a histogram, a bar chart or a gauge if these bring interesting insights to your audience and don’t crowd your map too much. Ask a colleague to take a minute to check it communicates what you want to say.

Source: CartONG demo data

As a general rule, stick to trusted data sources. Examples include the Humanitarian Data Exchange (HDX), ReliefWeb, government websites and recognised universities or research institutions. Make users aware of where the data is from, so that they can check it against the map, or use it themselves.

Have you decided to create a map with a dataset that you came across somewhere – but can’t remember where? Are the figures making sense? Whether or not you have any doubts, it is never a bad idea to triangulate your data.

The context of your map project will determine whether you should use absolute or relative figures. For Ebola and other diseases with extremely high mortality rates, absolute values may be more relevant, while for fast-spreading diseases such as COVID-19, relative figures might be more informative. For example, think about the different impacts of one hundred new cases on a population of one thousand or one million.

Source: The Guardian

In the context of an epidemic, it is often a great idea to show how the situation is evolving with a series of maps. However, if you would like to show a difference over time, make sure that the audience can see this very point – the difference. It is also important to show a meaningful difference. With very rapidly spreading diseases such as COVID-19, one or two new daily cases on top of the existing two hundred cases may not be statistically significant, while over twenty new cases accumulated over a week could start to make a significant difference to the situation.

Source: CartONG demo data

When taking a look at a map, you might think: “Why does this map say there are so few cases of COVID-19 yesterday? I heard in the news that there were at least 30 new cases, maybe many more”. It may be that crucial metadata is missing. Metadata is information about data, such as when the map was produced and when the underlying data was captured. With pandemics of any kind, data freshness is crucial for providing an accurate picture of the situation.

As well as dates, it’s important to share other information about the map. For example, do the administrative boundaries reflect the latest geopolitical situation? Who has provided the locations of the healthcare units? By including the metadata, you will be pre-answering many questions that the map’s intended audience may have about it.

Source: MapAction

Data on its own can be difficult to interpret without any context. Say, you are about to display 26 new and 465 existing cases in your country of interest. However, without further information about the context, it can be difficult for the audience to understand the full message of the map. Is the situation good or bad, deteriorating or improving?

In a context of a pandemic, you may want to compare to the situation elsewhere (e.g. other regions or countries). Another useful tip is to put the data into a temporal context by accompanying the map with a timeline chart. An alternative option to display this is to use a sequence of several maps – however, as mentioned in point 4, the whichever comparative element you choose, you should clearly display the difference between the two intervals.

Source: ft.com

In today’s world of crowdmapping and geotagging (with or without consent), it can be tempting to create maps in granular detail. However, you should always  consider medical data either at personal/household level or at community level as SENSITIVE information which requires a careful approach in terms of ethics, legal, organizational and technical protection measures. If you or someone in your family became infected with a disease, you probably would not like the whole world to see this on a map!

It is then important to raise questions such as : is it really appropriate or necessary for my operations to display data to such a detailed degree (e.g. at household or neighborhood/village level)? Are you absolutely certain that people cannot be identified through your aggregated data visualisation? Are you sure that there is no risk for the concerned population (e.g. mainly risk of stigmatisation and social threats in an epidemic context) to display such information?  Did you put in place appropriate measures to prevent people who don’t need to have access to the map or the source data (especially in case of dynamic maps based on line listing)? In case of doubt it’s always better not to share the map than to do harm!

Likewise, if your map is dealing with data obtained by surveys using website user tracking or similar, are you certain that respondents have consented to the mapping of their locations? If in doubt, double-check with the data provider.

Source: CartONG demo data

Most of the time, GIS practitioners are aiming to create products that are both informative and visually attractive. However, in rapid emergency response situations, maps often need to be ready in very tight deadlines. In this case, the priority is to create a simple map that shows the main point you wish to make. Cartographic enhancements such as colours, symbols and other possible data layers can usually wait. Therefore, before commencing the map building process, clarify the deadline and priority need.

An example of a simple map used in a context of flash floods at Djibouti.

Source: MapAction

Especially when working to tight deadlines, it can be tempting to cut corners with things like symbology, extent and labels. However, in a pandemic or epidemic situation in particular, it is wise to consider how the situation will evolve over time and factor that into your map’s design. This includes choosing a map extent, symbology and scale that can accommodate potential geographical expansion and a rapid or pronounced increase in cases.

• https://guides.mapaction.org/
• https://coronavirus.jhu.edu/map.html
• https://www.humanitarianresponse.info/en/infographics
• https://data.humdata.org/visualization/covid19-humanitarian-operations/
• https://covid19.who.int/?gclid=Cj0KCQjwqrb7BRDlARIsACwGad41c1zzFZ-Am2nM4lj43ETkBwr0mULaJ-0IxbxqF3Y1ph1QI4h7TJIaArgYEALw_wcB
• https://www.weforum.org/agenda/2020/08/covid-19-maps-tracking/
• https://coronavirus-resources.esri.com/

-----

This project was co-funded by the French Development Agency (AFD) and the H2H Network’s H2H Fund, the latter supported by UK aid from the UK government. Nevertheless, the ideas and opinions presented in this document do not necessarily represent those of the H2H Network, UK aid and AFD.