How are different African countries experiencing the impact of coronavirus? How do individual governments’ actions such as lockdowns play a role in containing the spread? Here I conduct a statistical analysis across several global regions.
Johns Hopkins University (JHU) Daily Confirmed Cases One of the leading health research institutions in the world. This dataset is one of the standard measures of reported corona cases across all countries. I am most interested in the virus’s impact on Africa, but for comparison I picked several large countries also affected by COVID-19. To make a standard comparison, I look at what happened in the X days since the first corona case.
Cumulative COVID-19 Cases Over Time Uganda took early, decisive action before there was a problem and not after. This has made an enormous difference — in this sample of 8 countries across the first 59 days of the virus in each country, Uganda has had the slowest spread.
Some countries (China, South Africa, Italy, U.S.) are growing too quickly with new virus cases and cannot be seen properly in the above graph. Here is the same data but with a logarithmic y-axis so that all series can be compared:
Unlike other countries, Uganda took action before the pandemic and not after it — this has made a world of difference.
Mobility The following chart illustrates mobility data from Google. Google has released aggregated, anonymized data collected from its Android and other devices across the globe which use the location feature. Google is able to tell which types of places are frequented by people more and less over time: retail and recreation, grocery and pharmacy, parks, transit stations, workplaces, and residential. I calculate the average decrease in “outside” activities — all the previous categories aside from residential. A value of -25 means, “On average, people go outside 25% less compared to pre-COVID times” for each country. This data is from Google and measures aggregate change in users’ location from February 15 — May 9, 2020.
In this sample of countries, Uganda has the biggest change in mobility of -50% and the U.S. has the smallest change of -15%. This may partially explain the drastically different rate of new COVID-19 cases in each country; when people go outside more, they of course come into contact with each other more frequently and end up spreading the virus. Interestingly, Italy has seen a significant decline in mobility yet the fastest initial rise in cases after China. Italy’s significant decrease in mobility started on March 8 when there were already 7,375 reported cases.
Mobility index over a 3-month period, constructed using Google’s data
Mobility index values as of May 9, 2020
Stringency Oxford University compiles a daily “stringency index” quantifying the different anti-COVID measures governments have put in place. This includes policies like school closures, movement restrictions, income support, COVID testing, healthcare investments, and others. Personally, I was surprised to see China having the lowest stringency index in this simple (as of May 17) because news articles paint a picture of very tough lockdowns there. In this stringency index, 0 means “not strict” and 100 means “extremely strict.”
Another interesting point is that Kenya has the strictest anti-COVID measures by this index, and it also reacted quickly to the onset of the virus, yet its rate of new cases has been faster than in Uganda and Rwanda. This may be due to the fact that Kenya’s mobility did not decrease as much as in the other sampled countries, or that it’s a country with greater levels of international trade and cross-border movements than landlocked countries like Uganda and Rwanda.
Oxford’s Stringency Index as of May 17, 2020
Disclaimers
I am not a medical expert and neither is Asaak a medical company. For the latest coronavirus advice, please consult your country’s health officials as well as reputable international health organizations.
JHU’s reported coronavirus case data depends on the accuracy of reporting per country. Most countries are facing shortages of COVID-19 tests. It is also possible that countries are consciously underreporting their confirmed cases.
Google’s data is limited by the low rate of smartphone ownership in Africa vs. other regions of the world. It is further limited by the relative sparsity of Google Maps locations in Africa vs. other regions, i.e. most parks, grocery stores etc. are usually not recorded in Google Maps.
All of the above mentioned data sources, particularly those within Africa, are limited and any correlations or relationships may not be predictive of the future. Furthermore, I have double checked my work but I still may have made mistakes in my calculations.
コメント