A Dashboard for evaluation & visualization of synthetic patient level data.

About

SYNDAT was developed as part of TA6.4 of the NFDI4Health Initiative. Main functionalities include:

• Automated, on-demand assesment of synthetic data quality & privacy metrics
• Visualization synthetic & real data relations using low-dimensional embedding plots
• Detection of possible outliers in the synthetic data population
• Visualization of distribution metrics in the form of violin, barchart or correlation plots

The Dashboard consists of a frontend module for user interaction and data visualization as well as a backend module for direct API access.

Installation

Python API

If you want to use the corresponding Python package that supports both evaluation metrics and visualizations programmatically, you can install it using:

pip install syndat

Documentation and Code are available from the following project: https://github.com/SCAI-BIO/syndat

Docker

You can run a local installation using docker-compose:

docker-compose -f docker-compose.local.yml up

After startup, you will find the frontend running on localhost:3000.

Running a local development version

Requirements

• Node.js 18+ LTS
• Python 3.x.x

# install python dependencies
pip install --requirement backend/requirements.txt

# start backend
cd backend && uvicorn api.routes:app --reload

# install node dependencies
cd frontend; npm install

# start frontend (Vite)
cd frontend && npm run dev

# optional: production build and preview
# cd frontend && npm run build && npm run preview

API authentication

The following two API endpoints for batch updating data as well as batch downloading data are secured by a basic authentication workflow:

• /datasets/import
• /datasets/export

The default username/password are defined in the backend environment file. You may change them before the application startup by adapting the corresponding system environment variables:

export SYNDAT_ADMIN_USERNAME=my_new_username
export SYNDAT_ADMIN_PASSWORD=my_new_password

Citation

If you use Syndat in your research, please cite as:

@article{Adams_On_the_fidelity_2025,
  author  = {Adams, Tim and Birkenbihl, Colin and Otte, Karen and
             Ng, Hwei Geok and Rieling, Jonas Adrian and
             Näher, Anatol-Fiete and Sax, Ulrich and
             Prasser, Fabian and Fröhlich, Holger},
  title   = {On the fidelity versus privacy and utility trade-off of synthetic patient data},
  journal = {iScience},
  volume  = {28},
  year    = {2025},
  doi     = {10.1016/j.isci.2025.112382}
}