Cloud

Access MIMIC on the Cloud

MIMIC-III and MIMIC-IV are made available via Physionet: MIMIC-III, MIMIC-IV. Beyond directly downloading the dataset from PhysioNet, there are a few mechanisms for accessing the data:

  • Accessing the data in BigQuery
  • Accessing the data on AWS (MIMIC-III)
  • Accessing the data in a Google Cloud storage bucket (MIMIC-IV)

We highly recommend using MIMIC in BigQuery for the following reasons:

  • No setup required
  • Updates will be integrated into BigQuery when they are available
  • Derived concepts from the MIMIC code repository are precomputed and available on the mimic_derived dataset

If absolutely necessary, it is possible to download the data from a Google cloud bucket; instructions are provided at the end of the document. Please do keep in mind that the PhysioNet team is covering the cost of downloading the dataset.

Accessing data on the cloud

There are three steps to accessing data on the cloud:

  1. Link your cloud account to your PhysioNet profile
  2. Request access to the cloud resource
  3. Log-in to the appropriate service and navigate to the resource

These steps assume you are already credentialed and have signed the data use agreement for MIMIC-IV. If you have not, read this page for instructions on gaining access to MIMIC.


Linking your cloud account

How to connect your PhysioNet account to a cloud account.

Accessing MIMIC-IV on the cloud

How to grant your linked cloud account access to MIMIC.

BigQuery

Access MIMIC IV on BigQuery.

AWS

Currently MIMIC-III is available on AWS but MIMIC-IV is not. MIMIC-IV will be released on AWS in the future.

Google Cloud Storage

Access the MIMIC datasets on Google Cloud Storage (GCS).