Hospital line data provision about covid

[wolverdude]

@hnykda we can just not show the line when the data is null.

Perhaps we should also put some sort of modifier on the line since countries are expanding their critical care capacity?

[lagerros]

I collected data on critical care beds per 100k population in various countries. The conversion needed to display it on our current graph is fairly straightforward, I think, and also included here. (So what's remaining is just taking the data from notion, displaying it in d3, and adding a note to the legend -- though make sure it follows/is incorporated into the new data format #161 and here) https://www.notion.so/9db3df0271214655b13c17bb7565c40c?v=5e75cc3aed784a6baf42826304e7a923

[wolverdude]

Should it just show up as a horizontal line like this?

It seems like this feature in particular will make zoom/scroll more necessary. Where is that in the pipeline?

I'm signing off for the night now. I'll look deeper into this in the morning.

[lagerros]

Yes, it should just be a horizontal line.

Zoom not currently in the pipeline; feel free to make an issue.

[wolverdude]

I think there's a piece of missing information here. The model is predicting active cases, but it isn't predicting cases requiring hospitalization. Presumably, these will be many fewer, and for any decisions this chart would help people make, that is the actual quantity they're concerned about.

To put the hospital line on this chart as-is would be inviting them to compare apples to oranges.

[lagerros]

The amount of active cases requiring hospitalisation is an empirically measurable quantity (20% IIRC), as is the amount requiring oxygen treatment (5% IIRC) (rough guesses likely fine for reasons we discussed on call this morning). So if you have x critical care beds you can support x/5%=20x active cases.

More details on a similar guesstimate in the notion database.

[lagerros]

[hnykda]

related: https://github.com/epidemics/epifor/issues/1

[hnykda]

@gavento in the new pipeline you made, this is not yet present, right? (I made it part of epidemics/epifor#8, but that's not part of the pipe).

Should we make it part of the estimate data?

[gavento]

@hnykda The main blocker is now the data itself - can someone data-savvy import it into the regions.yaml database, please?

Drawing plotly trace with single horizontal line is almost trivial.
We can also pass it in estimates or whereever (would not overthnink this, we may refactor into something later when we know more about our usage etc.)

[gavento]

Just noticed that #258 is a duplicate.

[hnykda]

I am closing this, done via #259

[hnykda]

followup #264