Inspiration
We were inspired by weather radar and heat maps, which give users a forecast of weather changes. We found that an easily accessible forecast for disease outbreaks would be a valuable application.
What it does
Under the Weather is a web application designed to forecast disease outbreaks in your area based on their inputed location (state). The application uses machine learning and data visualization to display possible outbreaks of either hepatitis, measles, pertussis, rubella, and smallpox in one of the 50 U.S. states.
How we built it
Frontend: HTML, CSS, TypeScript, React Database: MySQL Backend: Python, PyTorch, Plotly, NumPy, pandas
Challenges we ran into
- Project ideation
- Finding usable data
- Training the NN model
- Connecting the endpoints
- Time constraints
Accomplishments that we're proud of
- Record number of RedBulls consumed :)
- Carrying out a project from ideation to completion
- Trying out and learning new technologies/frameworks/languages
What we learned
- Developing/improving a neural network for predictive modeling in Python (using PyTorch).
- The basics of TypeScript and frontend development with React for the UI.
- Backend development and using databases with MySQL.
What's next for Under the Weather
- Data Collection: Identify additional infectious diseases (COVID-19, influenza, RSV) which are relevant to users and geographic region. Collect reliable data from public health organizations, research institutions, and government agencies.
- Real-Time Data: Incorporate real-time data with feedback from users (user location and possible infectious disease) to provide most up-to-date predictions.
- Qualitative Descriptions: Provide users with qualitative descriptions of predictions ("low risk", "medium risk", "high risk"), as well as suggestions for actions (wear a mask, avoid large group events, quarantine, etc).
- Mobile Optimization: Optimize application for mobile devices in order to improve accessibility.
Log in or sign up for Devpost to join the conversation.