Processing workflow

This document describes how the input CSV files are transformed into the derived patient-level summary data/derived_patient_status.csv in this demo repository.

Inputs

• data/sample_patients.csv — one row per patient with demographics.
• data/measurements.csv — repeated clinical measurements per patient (contains date, bp_systolic, bp_diastolic, weight_kg, etc.).

Script

The derivation is performed by the script:

• scripts/derive_patient_status.py

You can inspect or download the script and its runtime requirements here:

• scripts/derive_patient_status.py
• scripts/requirements.txt

This script:

• Loads data/sample_patients.csv and data/measurements.csv.
• Selects each patient’s latest measurement by date.
• Assumes an average height by sex and ethnicity (demo-only lookup).
• Computes BMI and derives weight_category and bp_category.
• Writes the one-row-per-patient output to data/derived_patient_status.csv.

Running locally

Install dependencies (if not already available):

python3 -m pip install -r scripts/requirements.txt

Run the script:

python3 scripts/derive_patient_status.py

This will write data/derived_patient_status.csv and print the number of rows written.

Notes

• The height assumptions and classification rules are demo-only and should not be used for clinical purposes.