Message Boards Message Boards

GROUPS:

Covid wave alert: statistical analysis and visualization

Posted 11 days ago
756 Views
|
2 Replies
|
13 Total Likes
|

enter image description here

Dear reader, in this notebook, we evaluate the situation regarding COVID-19 across the globe. We use the data provided by the Johns Hopkins University on github, for all countries. In the USA, we consider a special scenario in which every state can be counted as a different country, due to its large number of COVID-19 cases.

Since each country has a different way of accounting for its daily cases as a function of time, we need to smooth the data of daily cases. We take the difference between consecutive days, and also apply a filter to this quantity. As it is usual in Calculus, we make an analysis of the second derivative of the total number of cases per day, i.e., we evaluate the rate at which the daily cases increases or decreases; since the second wave is already happening in many countries, it is possible to evaluate and classify specifically the regions in which early increase rates, can be observed from its past rates distribution.

We classify all regions around the world within a scale: red (increasing the rate of daily cases), green (decreasing rate), and gray (non-classified zone due to a derivative being close to zero). We present this scale within a geographic map, and also as a function of time, for every week, starting by week 20. Countries are classified every week, showing their weekly status regarding the COVID-19 situation.

We thoroughly enjoyed the Hackathon and we wish this initiative helps authorities in taking early warnings as a way to prevent further waves.

Keep safe,

Antonio & Matheus