A Dashboard for evaluation & visualization of synthetic patient level data.
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
- Filtering of synthetic data & on-demand generation of synthetic data points
The Dashboard consists of a frontend module for user interaction and data visualization as well as a backend module for direct API access.
Login to your Github Docker Account:
docker login ghcr.io -u $GITHUB_USERNAME -p $GITHUB_TOKEN
Run the docker-compose
file in the root directory:
docker-compose up
Given all dependencies above are satisfied, both the frontend and backend can be started via docker-compse
with a single make command using the designated Makefile in the root dir:
make up
# 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 --legacy-peer-deps
# start frontend
cd frontend && npm start
You may alternatively start the backend and frontend servers using the make command make run-local
, given you have all required python dependencies installed already.
The following two API endpoints for batch upating data as well as batch dowloading data are secured by a basic authentification 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