ENCoDE_tutorial

A tutorial for use on ENCoDE, a skin tone and pulse oximetry dataset, can be found here: https://colab.research.google.com/drive/1YqEfUMU366IfpC1eaTnmMres45wd8O52?usp=sharing

Data Access

Data can be accessed through PhysioNet.org. Under the name: "ENCoDE, mEasuring skiN Color to correct pulse Oximetry DisparitiEs: skin tone and clinical data from a prospective trial on acute care patients".

Data Exploration

We have created a Docker image that contains characteristics and quality metrics for the ENCoDE dataset that can be visualized using the ARES web application. In order to run this image as a container on your local machine, you will need to:

1. If you don't have it already, install Docker
2. Execute the following command in your terminal to pull and launch the image:
   • docker run --rm -p 7070:80 -d --name ares ghcr.io/aiwonglab/ares-encode:latest (you may need to add sudo at the start of the command depending on your configuration)
3. Open a new tab in your web browser and navigate to localhost:7070 (or potentially, 127.0.0.1:7070, or 0.0.0.0:7070 depending on your OS) to access the web application and start exploring
4. When finished, run the following command to stop and remove the container:
   • docker stop ares

As we make updates to the dataset, we will plan to build their metadata into new versions of this same image; ARES allows you to explore and compare different data releases with regard to their quality and characteristics.