Mock Data Generator for COVID-MS Dataset

This repository contains a Python script to generate mock data that resembles the original COVID-MS dataset.

About the Original Dataset

The original dataset can be downloaded from PhysioNet. It contains various features related to COVID-19 cases among multiple sclerosis (MS) patients.

How to Use the Mock Data Generator without tunable wights

1. Download the original dataset from PhysioNet and place it in the same directory as the Python script (mock_data_generator_updated.py).
2. Run the Python script. It will prompt you to enter the number of records you'd like to generate.
3. The script will generate a CSV file (generated_mock_data.csv) containing the synthetic data.

How to Use the Mock Data Generator with tunable weights

1. Download the original dataset from PhysioNet and place it in the same directory as the Python script (mock_data_generator_with_weights.py).
2. Run the Python script. It will prompt you to enter the number of records you'd like to generate.
3. The script will generate a CSV file (generated_mock_data.csv) containing the synthetic data.

How to Edit the Weights

The script includes a dictionary called weights_dict that allows you to set custom weights for the occurrence of specific variables. Here's how it looks:

weights_dict = {
    "sex": {"female": 0.7, "male": 0.3},
    "covid19_admission_hospital": {"no": 0.8, "yes": 0.2},
    # ... other variables ...
}

To change the weights, simply edit the corresponding values in the dictionary.

How the Script Works

• The script first loads the original dataset to understand its structure and unique values for each feature.
• It then generates synthetic records, mimicking the data types and unique values found in the original dataset.
• Special care is taken to generate unique "secret_name" identifiers for each synthetic record.

Requirements

• Python 3.x
• pandas
• numpy

Disclaimer

This is a mock data generator and the generated data should not be used for any clinical or diagnostic purposes.