Lighthouse

Inspiration

One of the biggest barriers that people face when trying to access mental health is a lack of knowledge. It can be overwhelming to try and understand what type of treatment you'd benefit the most from, who's in your area that would be able to help you, etc, especially when there are an overwhelming amount of options out there. Lighthouse aims to be the first place people come to when they're lost while trying to get the help they need.

What it does

Lighthouse takes queries from a user based on what type of mental health support they're looking for, whether they're looking for a specific type of service, like CBT or group therapy, or for specialized support based on a part of their identity, such as support for LGBTQ+ people or survivors of abuse and domestic violence. Providers on the website are all verified prior to letting their profiles be shown by requiring proof of registration with a provincial certifying body for their occupation.

How we built it

We used Streamlit for the frontend, and FastAPI and Snowflake for the backend storage and query processing.

Challenges we ran into

One important part of our project was being able to use natural language queries, and a stretch goal was had was to implement a RAG LLM agent that would act as a guide to using the website. However, we ran into authentication issues with Snowflake and OAuth that prevented us from implementing those features, but we were still able to implement filter-based searches with important fields such as insurance and specialized support.

Accomplishments that we're proud of

Iman: This was Angie's first hackathon and she implemented most of the frontend, and did a great job with working with CSS and Streamlit! This was also my first time developing a backend and I'm glad I was able to implement the minimum features for our idea.

What we learned

Iman: I got to learn about working with Snowflake and backend technologies for the first time. Angie: I learned how to use CSS and streamlit!

What's next for Lighthouse

Implement a RAG LLM chatbot to answer users' questions and guide them towards what service may work best for them, based on their concerns Implement a navigation function that will give step by step instructions to users to help them get to a mental health care provider's office for in-person appointments (would be very helpful for people who struggle with navigating unfamiliar places) Provider self-registration and background check process Financial aid program where Lighthouse would be able to reimburse fees for the first few appointments for users who are unable to afford paid services, and partnerships with providers to allow for initial/intake sessions free of charge to allow users to see if they work well with a provider.

Built With

• fastapi
• snowflake
• streamlit