X-Hale | Devpost

Inspiration

With the rise of social media effects on mental health, we saw an opportunity to analyze these conversations and provide insights.

What it does

Our website leverages AI to detect trends in mental health-related posts and offer support when needed. Take a look at our quick presentation: https://docs.google.com/presentation/d/1XCIaSI9pyM3EXopcRVWOXL3PuyFbBm89B4isqSlZl5A/edit?usp=sharing

How we built it

To achieve this, we trained an AI model that find relevant posts based on keywords. The model then analyzes sentiment—classifying tweets as positive, neutral, or negative—to help understand public discourse around mental health. Check our repo for more details :) https://github.com/Cherryaki1/ConUHacksIX-Rice

Challenges we ran into

Our biggest challenge was the stability of our AI model—it crashed three times during training. However, the real setback came midway through the hackathon when we lost all our trained data and had to start from scratch. This was a major obstacle, but we managed to recover and retrain the model within the remaining time, which tested both our resilience and problem-solving skills.

Accomplishments that we're proud of

Despite the challenges, this project was an incredible learning experience. We walked away with new technical skills, a deeper understanding of AI and web development, and the ability to quickly adapt to unforeseen issues. This hackathon reinforced our ability to collaborate under pressure, and we’re excited to build on what we’ve learned in future projects.

What we learned

During this hackathon, we gained hands-on experience in training AI models and working with full-stack development using new technologies like Next.js and Flask. Additionally, we explored working with third-party APIs and integrating web scraping scripts to gather relevant data. These experiences pushed us to expand our skill set beyond our initial expectations.

What's next for Xhale

We hope to further optimize our AI model for better sentiment analysis accuracy, as well as improve the use of GenAI to offer appropriate references based on the trends analyzed.