Navigate to the root of the project and run:
pip install -r requirements.txt
This tutorial assumes you have a PosgreSQL instance with the MIMIC database running.
First run the query in mimic-data/mimic-sql/lab_measurements.sql
. This will give you a result set. Export it to CSV and name it mimic_lab_measurements.csv
.
Then run the script in mimic-data/build_query_for_features.py
like:
python build_query_for_features.py
This will generate a SQL file: mimic-data/mimic-sql/patient_features.sql
. Run this SQL file in your local PostgreSQL instance. Export its results to mimic_patient_featrures.csv
.
Then run the script in mimic-data/join_measurements_and_features.py
like:
python join_measurements_and_features.py
This will generate a CSV file: mimic_data.csv
. This is the MIMIC dataset used for training.