Inspiration

We realized that especially in rural areas, healthcare is something hard to come by. It can be hard to find a doctor near you or even figure out what sort of disease you may or may not have. So, we made it easier for the rural people to have proper medical care on time by solving their real life issues.

What it does

Health Warrior is a web application that allows you to sign in via Google and input your current symptoms. Next, our machine learning (ML) model will essentially predict what sort of disease you have and find the nearest doctor in your area. On your dashboard, you can view your history of the previous symptoms you submitted as well as the form that you can upload. Using geo location rural people can find nearby doctors and get themselves recommended to the respective doctor for their particular disease as well as symptoms.

How we built it

To create Health Warrior, we worked in a group with two back-end developers and two front-end developers. For the front-end development, we used HTML, CSS, and JavaScript for the code and pixlr to create and edit graphics. For back-end development, we created a ML model to predict a user’s disease and locate the nearest doctor. We used Python in combination with different databases and sponsored technologies including Radar.io. We used Flask, Google OAuth, to handle log ins then tracked each users previous historical data, in this data it gave them an estimation of what disease they had, with the percentages of each disease as well as the 3 nearest doctors to their location as well as how long drive it would take for them to get there.

We used the disease-symptom data, and built a Machine learning pipeline, which takes in the data from the user and gives a prediction, more specifically, it is able to give probability prediction of infection for each disease. Eg 1) : There is 25% chance that you have Cholera based on the current symptoms ... etc for all the diseases i.e ( 25% for Cholera , 30 % for Fever , 15% Migraine , 10 % No disease detected, 20% Flu ) We then integrated the model into Flask.

Challenges we ran into

One of main challenges faced was data collection for symptom and images of doctors. For the Geolocation and the machine learning, most of time was spent on data collection. We needed Images of Doctors, for this we used the GAN generated images.

Accomplishments that we're proud of

We are proud of the overall finished product we created. The UI and front-end development is smooth and consistent throughout the website, so we feel that the design we built makes our website user-friendly and easy to navigate. We are also very satisfied with how our back-end turned out and how we incorporated different databases and technology in our ML services to diagnose symptoms.

We are happy that we were also able to create something that solves a real world problem and that could be deployed to improve access to healthcare which is a very important issue, especially in the age of corona virus.

What we learned

We all learnt from one another about different programs and technology that we were not familiar with before. We worked in a team with people in different parts of the world and pulled an all-nighter to work on this project. With cooperation, teamwork, and utmost determination, we completed this project and implemented our idea by the deadline.

We learned that communication is important to create a finished and successful project and we also learned more about how to use machine learning and incorporate Radar.io technology.

What's next for Rural Health Care

We hope to use geo location for ambulance tracking in the future in the same way that we have used it to find nearby doctors. Learn more at Our Github repo, Our DevPost description and Our demo video

Share this project:

Updates