Geospatial data has long been used to track diseases. John Snow’s iconic mapping of cholera outbreaks in London in 1854 enabled him to identify the long misunderstood cause of the disease. Fast forward to 2014 and mapping tools and modelling were essential in the fight against the Ebola crisis in West Africa.
Today, geographic information systems (GIS) – which, simply put, offer a way to present spatial or geographic data in a meaningful way, be it through interactive maps or other infographics – have revolutionised this space. Hundreds of thousands of organisations now create billions of maps every day to tell stories and reveal patterns, trends, and relationships about almost any set of data you can imagine. The yearly global value of GIS and the related geospatial economy is estimated at $350bn.
When it comes to Covid-19, the GIS community has launched into action. From international bodies such as the World Health Organization and the Red Cross, to academics, local governments and hospitals, GIS is being employed both to powerfully communicate heavy data sets to the public, and to further analyse the numbers, eventually allowing for the creation of forecasting tools which can help policy makers prepare for what’s to come and allocate resources appropriately.
In a recent video on Youtube, Jack Dangermond, CEO of esri (the world’s biggest location intelligence company) who has pioneered the use of GIS software since the organisation’s birth in 1969, thanked the GIS community for its work during the pandemic. ‘Around the world, the GIS community is responding,’ he said. ‘It’s collaborating and sharing approaches and models and apps and data in various ways, and I’m very proud of this. The pattern that is emerging is one of building an information system for pandemics.’
The most common GIS app is the dashboard, used to display charts, maps, and other visual elements in almost real time. Dashboards are being used extensively to share information about Covid-19. The WHO’s dashboard, built from an ArcGIS platform, reflects its own data and is therefore a useful source for understanding official statistics. It also has a dashboard related to European cases.
Also phenomenally popular is the dashboard created by a small team at Johns Hopkins University in Baltimore. Updated in near real time, millions of people around the world have been using the map, which allows users to click on a country and identify confirmed cases and confirmed deaths. The underlying data comes from dozens of sources, including the US Centers for Disease Control and Prevention, the WHO, the European Centre for Disease Prevention and Control and the National Health Commission of the People’s Republic of China, as well as US state health departments and media-aggregating websites.
Another popular dashboard from the University of Virginia – called The COVID-19 Surveillance Dashboard – provides a visualisation of COVID-19 cases, recoveries, and deaths across the globe and is being constantly updated by the extremely busy team (they are currently on version 1.3.1). The dashboard includes a time slider which allows users to track the progress of the disease since 22 January 2020. It also includes a ‘download all’ button so that any interested party can get the latest complete dataset with one click.
Dawen Xie, a research scientist at Virginia’s Biocomplexity Institute, leads on the dashboard and has been working hard on it for the last few months (a team of eight, including PR support, keep it running). The dashboard is updated every two hours with new data, some of which is done automatically and some of which is checked and verified manually.
The latest iteration of the dashboard includes county-level rendering for the United States, and state/province-level statistics for China, India, Canada, Germany, Greece, Chile, Brazil, Argentina, Colombia and Peru. More countries are due to be added to this list next week.
Crucially, the dashboard is open to all and the team liaises with various policy makers who can use the data in their work. ‘We have different enquiries from federal agencies and different state agencies. They they want to use the data and try to get insight from it,’ says Xie.
From seemingly simple dashboards such as this can come very powerful tools. By integrating disease modelling into GIS systems, researchers can create prediction and forecasting systems. The team at Virginia are currently fine-tuning a new ‘medical resource demand dashboard’ which will allow users to identify where shortages in medical equipment are likely to reach crisis point.
‘We are using the forecasts from our epidemiological teams in order to try to forecast which areas are likely to run into resource crises,’ explains Mandy Wilson, also a research scientist at the Biocomplexity Institute. ‘So for example, if we have a surge of people who require hospitalisation, the hospitals may not have enough beds to accommodate that. And it’s an even more complicated problem when you consider that most people don’t go to the hospital and leave in the same day. So you could have a moderate surge in one week, but if most people with COVID-19 are staying for two weeks, then if you have even a moderate surge the following week you may already exceed your capabilities. And of course in New York, they’ve been having problems with not having enough ventilators and all the hospitals are running out of masks and other supplies. If you can predict when those surges are likely to happen, especially if you can flatten the peaks so that you can spread out the number of people who are coming to the hospital at the same time, then you can better plan.’
As well as predicting where crises might occur, such tools can enable users to forecast the outcome of different policy measures, such as social distancing. ‘We seed the model with infections as we observe them in the surveillance stream,’ explains research associate professor Bryan Lewis. ‘We'll make some assumptions about how many real infections in the population are represented by what we detect. Seems like those numbers are vary quite a bit. And then we just sort of churn the math – if one infection causes this many infections, if different age groups have different kinds of outcomes, etc etc. You can allow that to run forward under different scenarios, to offer projections about how policy decisions might change and push and flatten that curve.’
The process also works both ways, with GIS visualisations helping the forecasters whose models are plugged into the dashboards. Henning Mortveit is head of the software team at the Biocomplexity Institute. ‘This may not be such a glamorous thing to write about,’ he says, ‘but we also develop many, many pandemic models. We run many scenarios and the visuals that Dawen and Mandy can provide us with are really invaluable when it comes to calibration and validation of our models. It is so much easier to see. Does the model do something systematically wrong? We can we can visually see it right away. These kind of GIS tools support our core modelling and simulation tools.’
Ultimately, combining data, forecasts and models within GIS systems is all about making the information accessible and useable. Maps are being shared on large screens for emergency responders; map-driven apps are being used by first responders in the field. The Covid-19 pandemic is a powerful example of the importance of mapping and GIS in today’s integrated world. The huge amount of material being shared for free online is also a sign of an industry and community working together for the greater good.