Show HN: JSON API and Graphs for Coronavirus Cases | Hacker News Reader

Show HN: JSON API and Graphs for Coronavirus Cases

coronavirus.projectpage.app

26 pointsgrey-area16 comments

Thanks for sharing. I also started messing with a dashboard here: https://www.germanysbestkeptsecret.com/corona

Data I used is also from John Hopkins, turned into a JSON file via: https://github.com/pomber/covid19 .

There are also some good data-related resources here: http://open-source-covid-19.weileizeng.com

It's both disturbing and fascinating to work with this data. I would love to see these datasets go beyond the basics and also include things like timelines, population size, etc, to better understand how our actions influence the spread.

Looks really nice, thanks for sharing too, I like seeing the different approaches people take to this - I too had a large COVID-19 picture as a ground at one point! I love your little logo top left, did you make that?

I sorted by deaths as think it might give a more accurate picture than case counts - confirmed cases are really difficult to compare across countries give the different amounts and rates of testing.

Seems kind of crazy that a private university is the best source of data for this, there are still many inconsistencies in the data but it is a great resource and it's great they have released it to the public.

One thing I would like to fix in the current data is there is no breakdown of UK regions (there are UK dependencies oddly but none of the UK countries), and updates to UK data inevitably don't come till the US is awake. There are no global US, Canada, China datasets either in the data so those had to be created.

It'd be nice to see the WHO or someone step up and coordinate a database centrally - I think they have figures too but they are not very reliable or timely.

I found the icon on the Noun project, I'm attributing it a bit further down on the site.

Accuracy is super tricky here. Some people speculate that the numbers coming out of China are manipulated (I have no proof or opinion on that) and there's also some concern that deaths from other causes are also counted towards Coronavirus. But that's just what we have right now... an estimation of reality.

I'd also like regions, maybe that data is available somewhere, or could be scraped from national or local news sources, and then aggregated somewhere.

Are you planing to add more features, graphs, or other information?

I know regions are available for the UK on a daily basis (but not as a time series), so I'm planning adding that and recording the time series, though it does raise the question of inconsistencies in data between regional and national counts if I continue to use the national count from elsewhere.

I think I'll add testing figures as someone else suggested, though those are rapidly evolving and the latest figures I found were for March 19th.

This could be really useful, but seems it's not actually returning JSON, e.g. https://coronavirus.projectpage.app/.json?period=0

Oops, should be fixed now.

It's sad how Hong Kong is often left out or lumped together with China in these visualisations.

I'm glad at least Taiwan is listed separately though.

It would be great to also have the number of tests administered with this data set.

That is a variable that is adds to the relevance of confirmed cases.

I agree that'd be really good. I haven't seen any daily stats for that though, just isolated comparisons. [EDIT] - just found one (not daily, just single figures) and will add it later today.

Saw this tweet. Another interesting metric: Hospitalizations

https://twitter.com/natesilver538/status/1242208166202478593...

Related page with another API about Coronavirus: https://covid19api.com/ I haven't tried it yet, but it looks well designed. Data is sourced from Johns Hopkins.

There was a recent thread asking about this kind of dashboard:

https://news.ycombinator.com/item?id=22642370

Lots of very good resources there.

Thanks, I love this country comparator from your link:

https://boogheta.github.io/coronavirus-countries/#confirmed&...

Also this graph is a really good one from the FT of cumulative deaths by number of days from 10th death:

https://www.ft.com/coronavirus-latest

The FT one is since 100th case, not since 10th death, according to the graph.

There is a death trajectory graph (first graph) from the 10th death, and a case trajectory graph (second graph) from the 100th case.

The deaths graph is probably more accurate, since testing has varied so much.

nice but wouldn't the data be better formatted

[{ date: '', confirmed: 0, deaths: 0, recovered: 0}]

rather than sperate arrays.

Reply on news.ycombinator.com