search

jakobzhao / virus

Noval Coronavirus Infection Map
https://hgis.uw.edu/virus/
MIT License
161 stars 50 forks source link

Color coding by number of active cases, growth rate and growth factor? #16

Open airguru opened 4 years ago

airguru commented 4 years ago

As briefly mentioned in other issues, since some countries seem to be already recovering, It might be misleading to display them with darkest color in "infection level" view. So I would suggest to color them by number of active cases.

It seems like a simple enough task that I could do if you approve.

Also I think it would be good to add two new metrics (possibly with color coding option for both): 1) Growth rate - Percentage change in the number of confirmed cases (1-cases(today)/(cases(yesterday) - possibly an average over last 2-3 days? 2) Growth factor - new_cases(today)/new_cases(yesterday) - numbers around 1 are indicative of potential inflection point in the epidemics as described here: https://www.youtube.com/watch?v=Kas0tIxDvrg . Might be good to possibly also do at least 2-3 day average of this. Numbers below 1 are good, so I would color code them in green.

I believe I could do these as well.

Let me know what you think.

laurentlb commented 4 years ago

An option to divide the number of cases by the population of the country would also be interesting, as this changes the results quite a bit.

airguru commented 4 years ago

Well, that would be definitely a bit more complicated. Would you volunteer to gather the population data? :)

The question is, what exact message do you want to extract from that. The dynamics of the epidemic is captured well enough by absolute numbers (the "exponentialness") in my opinion. For instance, if there would be a consistent 25% day-on-day increase in reported cases in India, is it really a less severe situation when compared to the same absolute numbers and increases in less populous country? Maybe when judging the number of available hospital facilities?

I mean, I also got this thought, but I feel it has to be carefully considered.

laurentlb commented 4 years ago

This can be used to estimate the risk level? Some countries or companies require a quarantine when someone has visited a "high risk country". This table (https://www.worldometers.info/coronavirus/#countries) suggests that someone coming from Japan has lower risk of infection compared to someone from Norway.

jakobzhao commented 4 years ago

they are great ideas however we tended to be more cautious because a new index will also trigger alternative ways of interpretation.