Data for Singapore

[hxchua]

Hi! I'm looking to contribute data for Singapore as I've been collating and updating the data daily. I noticed the case count data required is [time, cases, deaths, hospitalized, ICU, recovered]. However, in the context of Singapore, we have people who are not hospitalized but are isolated and cared for at community facilities - as there's not enough beds in the hospitals to go around. These are those who have mild symptoms, or are clinically well but still test positive for COVID-19. They contribute to the count of total cases. How should I be including these people in the case count data?

The raw data can be found here.
More details about the dataset can be found here: https://data.world/hxchua/covid-19-singapore

Thank you!

[rneher]

thanks. we'll look into the Singapore data, but we would need a parser for these data and basic population statistics as described here: https://github.com/neherlab/covid19_scenarios/blob/master/data/README.md#contributing-and-curating-data

could you contribute such data and parsers?

[hxchua]

Yes, I can create the parser after your confirmation on what should be included in the hospitalised count. Will it be alright if I can only provide case count data but not population data? I noticed that the populationData.tsv dataset already contains Singapore. Thanks!

…

On Sat, May 23, 2020 at 4:45 PM Richard Neher ***@***.***> wrote: thanks. we'll look into the Singapore data, but we would need a parser for these data and basic population statistics as described here: https://github.com/neherlab/covid19_scenarios/blob/master/data/README.md#contributing-and-curating-data could you contribute such data and parsers? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#700 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AG2G3QABUZYF2VRNUHOJOY3RS6EKPANCNFSM4NIIYWFA> .

[rneher]

yes, you are right. we have Singapore data through the ecdc data set. Our model doesn't fit hospitalization data right now, but they are displayed on the graph. The models count for hospitalized cases is equivalent to those severely ill. So home isolation would not be part of this.

[hxchua]

Got it, thanks. Sorry another thing to confirm, is ICU part (i.e. subset) or independent of hospitalized?

[rneher]

We think of them as separate and more in terms of regular and ICU ward.

[hxchua]

When I tried to run the parser, I ran into this error: ModuleNotFoundError: No module named 'generated'.

In util.py, there's a line on import generated.types as schema. Appreciate advice on where does the module come from?

[ivan-aksamentov]

@hxchua Looks like you don't have the required Python classes that we generate from schemas (make types).
Typically, in order to run the parser step (download data and save to TSV files, make tsv-data) you don't need these classes. However, you need them in order to run the converter step (which generates the final JSON file that will be consumed by the application, make cases). Is it converter what you are trying to run?

Here is a quick guide:

We share type definitions between the data generators (Python) and the application itself (Typescript). In order to ensure strong typing across languages, we have JSON-schemas in schemas/ directory. There is a script tools/generateTypes.ts which takes these schemas and converts them to Python classes, Typescript typings and Javascript ajv validators. You need Node.js in order to run it. If you have Node.js you can run yarn schemas:totypes
from the project root, or run make types in data directory, which will do exactly the same:

covid19_scenarios/data/Makefile

Lines 11 to 12 in 5fdbf45

	types: clean-types
	cd .. && yarn install && yarn schema:totypes

After that the data/generated/types.py file will appear, and the things should start working.
Take a look at other commands in this Makefile if you want to generate the data for the application.