COVID-19 from a Data Science Perspective

The current figures and statistics on COVID-19 sometimes differ so widely that it is difficult to draw conclusions about actual infection and mortality rates. But in fact, these are the crucial factor for effective and appropriate measures. Daniel Haake is using all of his data scientific knowledge and comes to valid conclusions.

New numbers of COVID-19 infections and deaths appear daily. In addition, calculations on infection rates and mortality are published, that often seem to contradict each other in Germany alone. In an international comparison, the figures sometimes differ even more dramatically. So the question is: Which figures are correct and reliable? Or more precisely: How can representative studies provide more reliable figures and help to draw correct conclusions in order to take and assess measures?

Daniel Haake, M.Sc. in Data Science and Data Scientist at The unbelievable Machine Company, has spent a great deal of time and expertise trying to find the answers. With tremendous effort and attention to the details, he has gathered the facts and data on COVID-19 to date and evaluated them from a data science perspective.

To get an overview of the actual magnitude of the number of infections, he used representative studies. The COVID-19 cases on cruise ships (“Diamond Princess”) and aircraft carriers (“Charles de Gaulle” and “USS Theodore Roosevelt”) served as the first set of data, including an adjustment of the various median ages for transferring the results to an entire population.

In addition, he has used results of nationwide tests on randomly selected people in Iceland and a similar study by Stanford University in Santa Clara County, California. Both were designed to provide a representative view of the overall population.

The first interim results from Germany were provided by a comprehensive study in Gangelt (Heinsberg district in Germany), conducted by the University Hospital in Bonn, which Prof. Dr. Drosten has come to describe as an “extremely solid, robust study”.

In a revealing and continuously updated article, Daniel Haake explains what his findings signify for the officially reported cases in Germany. He explains what conclusions can be drawn about plausible and actual figures, the number of unreported cases and the lethality of those infected. Including urgent recommendations for action from a data science perspective.

[Read the full article]

This post is also available in: German