BioAcoustix

Inspiration

Funnily enough, our team was on the sustainability track because we originally planned on doing a project revolving around ecoacoustics and the ability to create a positive impact on the environment through guided urban planning. While brainstorming for potential backups, we came across BIOacoustics, which is similar in principle but much more geared towards life sciences, which turned us to the health and accessibility track.

What it does

Our project provides bioacoustic indicators to predict early diagnosis of Parkinson's as well as other diseases. Since the goal of the project is to provide an integrative service into telehealth platforms, the UX is made to be streamlined and all the user has to do is upload an audio clip to receive personalized diagnosis recommendations.

How we built it

Our technical stack consisted of React.js, HTML/CSS, and Python. We used Generative AI for audio denoising and chat features, APIs and datasets for various respiratory diseases, and IBM's Watson Assistant for Natural Language Understanding and Sentiment Analysis of Patient speech.

Challenges we ran into

Trying to stretch our project so that it dips into multiple tracks was difficult, especially when we struggled to find a good dataset and fine-tune our model. We also had ambitious goals of implementing real-time analysis of air quality/weather data and a clinician/physician login to view patient metrics, but we decided to focus on quality over quantity.

Accomplishments that we're proud of

Building a dynamic frontend is always a proud development feature. Alongside our frontend, we were particularly proud of the fine tuning we did for our model. We were also proud to commit upwards of 20 hours straight on this project as a group of 3.

What we learned

The most important lesson we learned was to choose a topic that the entire team is passionate about. We initially tried to push certain ideas because each of us individually had a strong interest or a strong background, but choosing a project that everyone was motivated to work on around the clock was definitely a key takeaway from this year's hackathon.