MIMIC-IV Demo Dataset substitute

[BPDanek]

Since access to MIMIC-IV sample data requires authorization from PhysioNet, for the purpose of the demo and example texts does it make sense to use this variant of the dataset instead? They are from the same author? https://physionet.org/content/mimic-iv-demo/2.2/

Pyhealth users may not be interested in the package if there are external barriers to using it.

[BPDanek]

To use the dataset, download it, and unzip it with archive utility (mac). Within the hosp directory you will find the .csv files which are supported for parsing into datasets. I had to install gzip (https://formulae.brew.sh/formula/gzip) to unpack these files since archive utility doesn't handle the .gz files by default. You can just unpack the files you'll use in your tables (in the example below it's the default tables and diagnosis_icd, procedures_icd, labevents)

dataset = MIMIC4Dataset(
        root=ROOT,
        tables=["diagnoses_icd", "procedures_icd", "labevents"],
        code_mapping=CODE_MAPPING,
        dev=DEV,
        refresh_cache=False
    )

[BPDanek]

Since this is a demo dataset, you may be able to put it behind one of the googleapi links and allow really easy usage

[ycq091044]

Done! We already linked to the demo dataset https://physionet.org/content/mimic-iv-demo/2.2/ and uploaded it to googleapi.