Spark PrescrAid

We propose a system, in which medicines’ reactions are reported, and a Spark based artificial intelligence tool aids medical doctors in prescribing the best medicines for a given set of ailments.

Comment

Inspiration

Covid-19 pandemic has warned us that the regulatory system for medicines should not stop at the approval and release. We genuinely believe that Apache Spark can help us improve this system.

What it does

This project proposes a system, in which medicines’ reactions are reported, and a Spark based artificial intelligence tool aids medical doctors in prescribing the best medicines for a given set of ailments.

We propose web sites, where people can report reactions of various medicines in the presence of various health conditions.

Reports will be accumulated on a server, which will feed the data into a Spark based machine learning algorithm. The streaming data will develop and update models for medicine appropriateness.

How we built it

We realized that doctors need a tool to help them in finding the most appropriate medicine(s) from the currently complex and confusing zoo of medicines.

For the front-end part of the project, we set up an illustrative web site at lipy.us/sparkai and preaid.net/sparkai. In view of the availability, we have used an Apache web server with php scripts rather than Node js web server.

The back-end part of the project is taking shape. We are using Apache Spark, with coding done mainly in Scala. We are using Spark Streaming with appropriate machine learning libraries.

To keep the cost low, we are using our own desktop rather than some cloud services.

Challenges we ran into

Dreaming is easier than doing. We had some trouble, and fun, in making some libraries work.

Accomplishments that we're proud of

We have made a reasonably good front-end so far. We have made a good beginning for the back-end.

What we learned

We learned a lot while dealing with Spark Streaming and Spark's ML libraries.

What's next for Spark PrescrAid

We will make the "Eagle Land"!

Thanks

We thank the Open Source ecosystem for its amazing tools and techniques, Databricks for sponsoring this hackathon, and many fellow participants for inspiring us in various ways.

Built With

Share this project:

Updates