Plot rates by default in addition to numbers?

[poldrack]

there are various ways in which one might normalize the count data that could be relevant to understanding and prediction. for example, per capita, per # of hosiptal beds, per square mile, etc.

[GaelVaroquaux]

There are two aspects:

• Normalize per capita, ... related to #3 (which could have been better worded)
• Growth rate. These cannot be estimated with a naive model, as they need to account for active cases, and not only total confirmed. I tried yesterday evening, failed, and need to go back to it.

[GaelVaroquaux]

@emmanuelle : you had an idea on where to get the data (population of each country) for this, I think.

[GaelVaroquaux]

I think that this will do the trick: https://download.geonames.org/export/dump/countryInfo.txt

[emmanuelle]

There is https://population.un.org/wpp/

[emmanuelle]

and https://www.gapminder.org/data/

[GaelVaroquaux]

I've merged #39 which normalizes the map by population.

I am wondering: should we also normalize the time series? Unlike with the map, we cannot keep all the information to have the un-normalized version in the hover (too much to store). @poldrack : any opinion?

Also, @poldrack : any advice on how to name the corresponding quantity in a way that is easy to understand for the non specialist? I wrote "Active cases per Million". I'm not sure that this is good English.

[poldrack]

I'm fine with leaving the timeseries un-normalized. could be useful to normalize it but not essential. I think your wording of "active cases per million" is fine, assuming that this is actually what the data refer to. is that what they are called in the JHU dataset?

[GaelVaroquaux]

is that what they are called in the JHU dataset?

The JHU dataset that I have doesn't give this. I computed them myself.

Thanks!

[poldrack]

can you tell me exactly how?

On Tue, Mar 17, 2020 at 11:58 AM Gael Varoquaux notifications@github.com wrote:

is that what they are called in the JHU dataset?

The JHU dataset that I have doesn't give this. I computed them myself.

Thanks!

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/covid19-dash/covid-dashboard/issues/8#issuecomment-600243082, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAGUVEHRPUAKF6YH4PBUK53RH7B5FANCNFSM4LJUBKEA .

-- Russell A. Poldrack Albert Ray Lang Professor of Psychology Professor (by courtesy) of Computer Science Bldg. 420, Jordan Hall Stanford University Stanford, CA 94305

poldrack@stanford.edu http://www.poldracklab.org/

[GaelVaroquaux]

The relevant lines are: https://github.com/covid19-dash/covid-dashboard/blob/master/make_figures.py#L29

[poldrack]

so you are not removing recovered cases and deaths, correct? in that case, the more appropriate term would be "cumulative cases"

On Tue, Mar 17, 2020 at 12:07 PM Gael Varoquaux notifications@github.com wrote:

The relevant lines are:

https://github.com/covid19-dash/covid-dashboard/blob/master/make_figures.py#L29

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/covid19-dash/covid-dashboard/issues/8#issuecomment-600246837, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAGUVECRU2YW4TC4IYVVOXLRH7C5ZANCNFSM4LJUBKEA .

-- Russell A. Poldrack Albert Ray Lang Professor of Psychology Professor (by courtesy) of Computer Science Bldg. 420, Jordan Hall Stanford University Stanford, CA 94305

poldrack@stanford.edu http://www.poldracklab.org/

[GaelVaroquaux]

so you are not removing recovered cases and deaths, correct? in that case, the more appropriate term would be "cumulative cases"

Everything that we are plotting is "active": removing recovered and fatalities: https://github.com/covid19-dash/covid-dashboard/blob/master/data_input.py#L40

[poldrack]

ah, sorry I missed that - so yes, "active cases" is fine then

On Tue, Mar 17, 2020 at 12:19 PM Gael Varoquaux notifications@github.com wrote:

so you are not removing recovered cases and deaths, correct? in that case, the more appropriate term would be "cumulative cases"

Everything that we are plotting is "active": removing recovered and fatalities:

https://github.com/covid19-dash/covid-dashboard/blob/master/data_input.py#L40

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/covid19-dash/covid-dashboard/issues/8#issuecomment-600252005, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAGUVEAMVFF2SS3KSHCKOJLRH7EKRANCNFSM4LJUBKEA .

-- Russell A. Poldrack Albert Ray Lang Professor of Psychology Professor (by courtesy) of Computer Science Bldg. 420, Jordan Hall Stanford University Stanford, CA 94305

poldrack@stanford.edu http://www.poldracklab.org/

[GaelVaroquaux]

I think that this is fixed. Closing.

[GaelVaroquaux]

Rethinking about this, I think that one of the main way people use this site is to compare countries. Hence, I am reopening, and I am in favor of plotting the normalized data in the time series plot.

We cannot display both information, because I fear that it would make the website slower, and it is currently quite slow: there is a big delay between clicking to select a country and the update. This delay makes using the website awkward.

What do people think? @poldrack @emmanuelle

[poldrack]

yes, I think that cases per capita is much more interpretable and relevant to things like hospital capacity

On Wed, Mar 25, 2020 at 7:42 AM Gael Varoquaux notifications@github.com wrote:

Rethinking about this, I think that one of the main way people use this site is to compare countries. Hence, I am reopening, and I am in favor of plotting the normalized data in the time series plot.

We cannot display both information, because I fear that it would make the website slower, and it is currently quite slow: there is a big delay between clicking to select a country and the update. This delay makes using the website awkward.

What do people think? @poldrack https://github.com/poldrack @emmanuelle https://github.com/emmanuelle

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/covid19-dash/covid-dashboard/issues/8#issuecomment-603878878, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAGUVEGPOV5SDYUPZV76FSDRJIJ63ANCNFSM4LJUBKEA .

-- Russell A. Poldrack Albert Ray Lang Professor of Psychology Professor (by courtesy) of Computer Science Bldg. 420, Jordan Hall Stanford University Stanford, CA 94305

poldrack@stanford.edu http://www.poldracklab.org/

[GaelVaroquaux]

OK, I'm on it.