The initial inspiration was the desire to improve public cleanliness and make it easier for people to report and track issues in their community. Additionally, incorporating maps into the system could allow for a more visual and user-friendly way of displaying and navigating the data.
What it does
A web app to create a "germ map" of various public places that highlights areas with high risks of germs and general bad hygiene.
How we built it
To build such a system, we used a combination of technologies such as Google Maps API, data storage and Machine Learning / AI algorithms, and a user interface for submitting and viewing reviews.
Challenges we ran into
- Accurately geolocating user-submitted reviews, handling large amounts of data, and ensuring the system is user-friendly and accessible for all.
- Setting up logs for ML experimentation was challenging.
- Had to constantly reshape input data for different models. ## Accomplishments that we're proud of
- Trained two ML models for regression and clustering with multiple experimentations within 24 hours
- Built a fully functioning cross-platform frontend in a time crunch
- Successfully built a statistical model and highlighted clusters of data on a geolocation interface ## What we learned
- Expand our skill sets in full stack development through Flask as a backend and Flutter as a functioning frontend.
- Importance of user-centered design, and how to design a system that is easy and intuitive for people to use
- Issues related to public cleanliness, and the importance of community engagement in addressing these issues. ## What's next for BacTrac
- Gain more accurate insights through geotagging
- Implement NLP models to better analyze user reviews and make it more accessible.
- Add support for multiple languages
- Reinforcement learning to update forecast in real time
- Try and implement anomaly detection to accurately predict outbreaks in association with public health and CDC datasets