MEDEC

A website that provides basic medical information and diagnoses for people without affordable healthcare.

Comment

Inspiration

Both of us come from India, a country where medical issues that go undiagnosed is a common issue. As well as this we were startled by the statistics on the sheer amount of people without access to healthcare, and sought to try and alleviate the problem. By building a website and combining our skills we were able to do just that. We were able to bring healthcare and information to those who need it most.

What it does

Our site provides medical information to those without affordable healthcare. It diagnoses a person based on their entered symptoms and gives them information on those diseases. The goal of this is to inform people on when exactly they should contact a doctor so that they contact them in time. It takes in 3 or more symptoms and outputs 4 predicted diseases through a ML model. Those diseases are also ranked by percentage on the likeliness that you have that disease.

How we built it

We used machine learning for the predictions, along with Python(Flask), JavaScript, HTML, and CSS. The back-end uses a neural network with lots of data to predict the disease and determine the percentages. The python application uses flask to host the website as well as host the ML model. We then have the HTML and CSS pages that actually have the design of the pages. We use JS to get the information that the user enters and then send it back to the python file. The python file then runs the data through the ML model and sends it back to the JS to then display on the website.

Challenges we ran into

We had challenges with the neural network model's accuracy as well as connecting it to the front end. This was the biggest project we have ever done, spanning 4 languages, and may long hours. Connecting the front-end and back-end was a new concept to us and the ML model didn't make things easy. We had trouble relaying information from the website back to the python and ML files. We also had lots of trouble getting the model as accurate as possible. In fact adding the iframe of the CDC website took lots of work as well. We also unfortunately couldn't figure out how to make the website adapt to different resolutions, and we hope to fix that later.

Accomplishments that we're proud of

We are proud that we were able to get both the model and the Flask API finished as those are new concepts to us. We learned so much through this project about training models, connecting different languages as well as user experience. We're proud that we have a finished product, and moreover we're proud that we built something that can help people.

What we learned

We learned about the application of AI in an actual industry. We have never done a project like this before and we sought to actually help people through our app. We learned what it means to make an actual website with HTML and CSS. Many of these languages we'd never used before, but we got a chance to learn this time. We learned a lot about the flask API and how ML processing works with actual input from a user.

What's next for MEDEC

We hope to improve the model and maybe even release it as an actual website. We're very proud of this project, and we hope to take it further. With better and more data, our model can reach higher accuracies and span wider symptoms and illnesses. We really want to continue this project because we know that it has the potential to help millons.

Share this project:

Updates