COVID-19: Visualizing and analyzing early trends

For my Data Visualization class I made a poster to visualize the trends shown by the of existence of public testing and positive case data for several countries at the beginning of the COVID-19 pandemic (until May 13, 2020).

At this stage, there were many unknowns as to what good public policy entailed when dealing with the COVID-19 pandemic. In order to distill a preliminary "evaluation" for the performance of each country hit by the pandemic, I proposed the curve formed by the ratio of tests performed over confirmed COVID-19 cases.

(In Spanish) Explanation of the movements in the curve.

When the curve grows, it is either due to more tests being performed (indicating more reliable data) or less cases detected (indicating that less people are being infected).

On the other hand, when the curve decreases, it means that the country will have a harder time dealing with the pandemic, either due to having less information to analyze, or because cases are rising.

(In Spanish) Example for South Korea's curve.

The leftmost arrow shows that a peak in cases without an equal increase in testing will cause the curve to decrease, therefore indicating that the state of the pandemic (in these early stages) is worse for the country due to cases increasing faster than tests do.

The rightmost arrow shows that when cases remain mostly constant and tests raise faster than positive COVID-19 cases, the curve will rise steadily. This shows a relative improvement in the country's outlook for the pandemic.

(In Spanish) Comparison between countries and territories with accessible public data regarding number of tests applied per day along with number of positive cases per day.

This figure shows the smoothed (by a 5 day moving average) curve of tests over cases for each country presented. This visualization allows the reader to quickly compare several countries. A sustained blue over several days is best, as it shows that the country is either improving testing beyond its raise in cases (thus giving it more information to make decisions with), or it shows a drop in cases (leading to controlling the pandemic).

On the other hand, deep red ranges show a country's capabilities being exceeded by a lack of testing or a raise in cases.

In addition to the smoothed curves, I also present several measures relating to each country: Human Development Index, Democracy Index, and Corruption Perception Index. These are presented in order to quickly compare whether factors such as corruption, authoritarianism, or human development actually affect the capability of a country to respond to COVID-19 in its early stages.

This poster shows an initial approach at measuring how a country can respond to a novel pandemic when limited resources and information are taken into account. While we now understand the COVID-19 pandemic far better, I strongly believe that this approach to analyzing highly uncertain data gave insights as to which country would fare better than others in the beginning and middle stages of the pandemic.

Some of the observations made in the document were:

• It is normal that any given country will have a steep decrease in its curve at first.

• In general, a country with a lower corruption perception will manage to increase its curve by the first 2 weeks.

• Authoritarian countries don't seem to fare better than others in general.

• Asia-Pacific countries had a tendency to grow their curves, while Latin American countries (except Cuba) didn't.

• For the period observed, the best performing countries included Vietnam, Thailand, South Korea, New Zealand, Malaysia, Italy, Denmark and Austria, while the worst performing were Russia, Belarus, Mexico and Ghana.

The full poster can be viewed here.

The data and methodology used can be viewed here.

You may also like