schoolofdata-ch / swiss-hospital-data Goto Github PK

In our project we have started from a single source, but are expanding to additional ones. It's becoming clear that there will be likely one or more sources per region, in some cases even per medical facility. Therefore the list that we are currently tracking in our ideas notebook should be formalized. The Data Package spec provides a good way to do this:

"The raw sources for this data package. It MUST be an array of Source objects. Each Source object MUST have a title and MAY have path and/or email properties. Example:"

"sources": [{
  "title": "World Bank and OECD",
  "path": "http://data.worldbank.org/indicator/NY.GDP.MKTP.CD"
}]

• title: title of the source (e.g. document or organization name)
• path: A url-or-path string, that is a fully qualified HTTP address, or a relative POSIX path (see the url-or-path definition in Data Resource for details).
• email: An email address

See also: neherlab/covid19_scenarios_data#3

We need to set up a simple process to ensure reproducible results. I would use DataFlows, but Snakemake, Renku, Nextflow and Workbench have been suggested by our team as well. I think we should at least stay true to DataFlows philosophy:

• use it for data that easily fits on your hard drive
• all data produced by these flows is easily reusable by others
• not a heavy-weight framework
• machine and human readable interface

Since starting this project we have debated a lot about what kind of information we need and don't need. It is clear that our starting point, the available Open Government Data, has both too much and too little information. We want to expand and improve this schema, and write a converter to stay compatible in case there are future updates.

We are working on a data dictionary that is complementary to the schema, and explains the fields in clearly accessible terms. Currently we are especially interested in these indicators, which need to be clearly explained and defined:

    N of hospitals in CH (with name and geolocation)
    N of beds by hospital
    N of ICU beds by hospital
    N of respirators by hospital
    average % bed availability pre-COVID-19 (“Bettenauslastung”) by hospital
    N ICU nurses by hospital (“IPS Pflege”)

A repository was also started yesterday with an example dataset which aims to simplify and standardize the requests we are making to regional health authorities: https://github.com/openZH/covid19_hospitalinfra

Our current data package has an automatically generated schema, which can be edited in the datapackage.json file, or loaded in the Frictionless Data Creator. The results of the data dictionary and schema proposal work can be merged in this way.

We need someone to digitize the data on ICU beds from here:

https://www.swiss-icu-cert.ch/files/daten/Dokumente/liste_zertifizierte/Internet_Liste_SGI_zertifizierte_Intensivstationen_200311.01.pd