search

jeremymoreau / covid19mtl

COVID-19 epidemiological dashboard for Montreal
https://covid19mtl.ca/en
MIT License
30 stars 11 forks source link

Add figure showing effective reproduction number (Rt) #53

Open jeremymoreau opened 3 years ago

jeremymoreau commented 3 years ago

Data from the INSPQ here: https://www.inspq.qc.ca/covid-19/donnees/rt

A similar styling to this website would look nice (i.e. red above 1, green below 1): https://rt.live

There is an option to download a csv, but I can't seem to find the URL (maybe dynamically generated?) rt csv

mschoettle commented 3 years ago

Great idea! INSPQ's numbers end Dec 11 right now so they seem quite delayed. But it seems to be available as a CSV here: https://www.inspq.qc.ca/sites/default/files/covid/donnees/rt.csv

There is also this here: https://epiforecasts.io/covid/posts/national/canada/ which has the calculated QC Rt. The result is similar to Ryan Imgrund's calculation who shares it on Twitter (https://twitter.com/imgrund/status/1343196458514833408). He also has one for Montreal. I assume the calculation of Rt is not trivial to do it ourselves, haven't looked into that though.

jeremymoreau commented 3 years ago

Thanks, that's super useful! The INSPQ says they update Rt every Monday and show data up to 10 days ago (I see the numbers going up to Dec 18 now, so seems like they updated it today). The other link you shared also gives us partial data estimates and forecasts, which is nice. Seems like we might need to use their API to get the data though as the URL changes for every version of the dataset: https://dataverse.harvard.edu/file.xhtml?fileId=4278773&version=60.0.

And yes agreed, I wouldn't trust myself to calculate the Rt as there are many possible pitfalls and I'm not an expert on this. I think I'll start by adding the INSPQ data since that seems easier (although it would be nice to have the forecasts as well).

mschoettle commented 3 years ago

Sure, we can always switch to the other one after.

I did find their GitHub repo (https://github.com/epiforecasts/covid-rt-estimates) which contains a link to the CSV (https://dataverse.harvard.edu/api/access/datafile/4159130?format=original&gbrecs=true). ~This might be to the latest version~ It is unfortunately not. But the repo also contains the CSV so we could pull it from there directly. Otherwise, the API has a way to get the latest version of a document (see: https://guides.dataverse.org/en/latest/api/dataaccess.html#download-by-dataset-api).

jeremymoreau commented 3 years ago

Thanks for finding that! Yes, maybe let's use the epiforecasts data then since they provide up-to-date forecasts. I think we can just grab this csv everyday? https://raw.githubusercontent.com/epiforecasts/covid-rt-estimates/master/subnational/canada/cases/summary/rt.csv