djay / covidthailand

Thailand Covid testing and case data gathered and combined from various sources for others to download or view
126 stars 15 forks source link

7 day rolling average as % of peak #87

Closed djay closed 2 years ago

djay commented 2 years ago

as on https://graphics.reuters.com/world-coronavirus-tracker-and-maps/vaccination-rollout-and-access/ Screen Shot 2021-09-30 at 12 37 25 pm

original reason for the infections esitmate as a fairish way of showing the correlation between the deaths and cases. this is perhaps a better way of doing that. I think one plot could show

Tests and cases should help show that it's not really the less you test the less cases you get. The rest should help so if vaccinations are working (less severe cases per cases), or if testing is working (deaths/severe isn't more than cases).

Similar plot could be used to show mobility vs cases?

djay commented 2 years ago

also the infections estimate probably needs to be taken down. IFR becomes less and less accurate due to vaccinations and was probably not accurate with delta anyway.

reduxionist commented 2 years ago

woops referenced wrong issue in that unrelated commit.

anyway, just note to self: salving this issue would probably render #35 moot too.

reduxionist commented 2 years ago

My notes based on guidance from @djay on how to implement this:

Add to covid_plot.plot_area

In func args

          normalise_to_peak: bool = False

In func doc string

    :param normalise_to_peak: whether to normalize data against peak value

In func body

    if normalise_to_peak:
        df = df / df.loc[time_period].max()
reduxionist commented 2 years ago

I am moving along to #51 for now, so anyone who wants to dive in -- please feel free! :smiley: