Hi folks,
Hope everyone is doing well. A colleague of mine ( Christian Schmid) proposed a simple way to estimate the Reproduction Number for COVID-19 using the historical date of onset of cases. The idea is to choose a single day that someone is most infectious and assume that all new cases on that day were due to the actively infectious people. It is a major simplification, but makes it very easy to model. If we perform some smoothing to the actual data, it turns out to provide results that pass the smell test (at least to me).
So all we need to do is (1) smooth the case counts using a WINDOW_AVG function, and (2) calculate the ratio of infectious to newly infected using a LOOKUP function.
Attached is a Tableau workbook that you can reuse. Simply modify the spreadsheet a put the count of cases per day in for your region. Make sure to put the date of Onset, not detection/test. Of course, you can swap out the Excel file for another data source if you have the data elsewhere. Use the two parameters to adjust the amount of smoothing and the day of peak infection.
Alberta has possibly dropped below 1 (meaning every new infection is only infecting 0.8 others!?). With the social distancing being relaxed, will the R value rebound? To what value?
