DRAVA: Utilizing Disentangled Representation Learning as A Visual Analytics Method for Pattern-based Data Exploration

This is the source code for our CHI 2023 paper, DRAVA: Aligning Human Concepts with Machine Learning Latent Dimensions for the Visual Exploration of Small Multiples

This repository has two main components: a frontend interface and a back-end server.

Development

Please refer to the documentation webpage for more details

Run the doc webpage

docsify serve docs

Backend

The backend is developed and tested with python@3.7.9

First, install all dependent packages:

cd server
pip install -r requirements.txt

Then, start the flask server:

cd flask_server
python app.py

To manage dependencies more effectively, you can create and use a conda environment before installing all packages:

conda create -n drava
conda activate drava
# ...
conda deactivate

The pre-trained models are stored at server/flask_server/saved_models.

Frontend

First, install all dependent packages:

cd front
npm install

Then, launch the Drava react application on the browser:

npm start

Required Datasets

To run both the server and the client, you need to put additional files into your local repository. These include

• server/data/ (Image patches of JPG files and compressed numpy arrays of .npz files)
• front/src/assets/ (JSON files that specify genomic ranges)
• front/public/assets/ (CSV files that contain external analysis results)

These datasets are shared upon request.