AI-powered COVID-19 Watch 

Methodological note

10 AI-powered COVID-19 Watch 

OECD.AI’s AI-powered COVID-19 tracker provides up to date information and news about the coronavirus. Interactive visualisations are made available to policy makers and other users to help them track and compare the evolution of the pandemic between countries.

The tracker’s main dashboard consists of two panels (Figure 1). The left-hand panel contains a table where is country is paired with up-to-date national COVID-19 statistics, including infections, deaths, and recoveries. Dynamic visualisations are accessible through the different tabs available in the right-hand panel. For each tab, it is possible to compare the evolution of the pandemic in a selection of countries by clicking on them in the left-hand panel.

Figure 1. AI-powered COVID-19 tracker 

Sample dashboard illustration

Live News tab 

The first tab displays live COVID-19-related news from around the world. The feed comes from Event Registry (see Event Registry data) and uses “COVID-19” and “Coronavirus” as the two keywords upon which news are selected. Country-specific views are available by selecting a country from the list.

Evolution tab 

This visualisation displays the evolution of the pandemic though five dimensions: the number of infections (x-axis); the growth rate of infections (y-axis) (the growth rate of infections is estimated using an exponential regression model based on the number of infections in the previous 7 days); the number of deaths (bubble size); the region (bubble colour); and the date (timeline). 

In addition, a red ring around the country’s bubble indicates the date when the first COVID-19 death was reported. 

By checking the “Switch deaths/cases” box, the user can swap the x and y-axes to display the curve of deaths instead of infections. The “Cases per capita” and the “Deaths per capita” checkboxes allow the user to normalise the numbers per unit of population.

To spread the long tail of the distribution and thus allow for a more intuitive visualisation, the square root of the x-axis is set as the default option. However, a checkbox makes it possible to display the data in linear scale.

When a specific country is selected from the list, the country-tracking mode is enabled (Figure 2). While still displayed, all other countries are shaded and a trail is drawn for the country of interest. The user can decide to “Hide other countries” and/or “Hide trace” by clicking on the relevant buttons. The user can select multiple countries from the list by holding the Ctrl key down.

Figure 2. COVID-19 “Evolution” tab 

Sample illustration of the evolution of the pandemic in the EU27 

Progression tab 

This line chart displays the number of infections or deaths – either the daily amounts or the cumulative number from the start of the pandemic – by country and in time. Although the default x-axis scale is linear, the user has the option to change it to logarithmic to better visualise countries with large differences in value. The top 10 countries by the number of infections or deaths are shown by default, which could be customised by selecting different countries from the list (Figure 3).

Figure 3. COVID-19 “Progression” tab 

Sample illustration of the progression of the top 10 countries by number of infections or deaths 

Trajectory tab 

While the “Progress” tab displays the number of infections or deaths in chronological order, the “Trajectory” tab sets a unique starting point for all countries as the base for comparison. Here, the x-axis measures the number of days after a given country reached 100 infections. The goal is to allow for fairer comparisons of different countries’ trajectories, as if they all started fighting the pandemic on the same day. The starting point has been arbitrarily set to 100 infections as an indication of COVID-19 spread capacity in a given country. As for the “Progression” tab, the scale of the x-axis could be switched to logarithmic and the country selection can be customised by selecting different countries from the list (Figure 4).

Figure 4. COVID-19 “Trajectories” tab 

Sample illustration of the trajectories of the top 10 countries by number of infections or deaths 

Worldometer data tab 

This tab presents a table with live COVID-19 statistics, including infections, deaths, recoveries, and tests. The data are retrieved from Worldometer and are updated several times per day.

Data Sources 

Except for the Worldometer data tab – which retrieves its data directly from Worldometer – the AI-powered COVID-19 watch on OECD.AI uses official data from the World Health Organization (WHO), which is retrieved daily from the John Hopkins University’s GitHub repository. This is the data repository used for the 2019 novel coronavirus visual dashboard operated by the Johns Hopkins University Center for Systems Science and Engineering (JHU CSSE).