Daily Journal

• 

  website

Inspiration

Mental health is a big issues and having a way to have people get help is always a positive. In the past we have seen research into how NLP can be used to find possible diagnose in mental health. The idea is that if we create an app that encourage people to write they may find it therapeutic.

What it does

A user can build a journal on a website. The journal when submitted will get a score from our LLM. Then it will be stored on mongoDB. In addition we also created a random forest model that predicts what demographics of students who mental health is most at risk.

How we built it

We used react for for the frontend and flask for the backend, and we stored journal data in mongodb. We are also running the ML models on the flask server.

Challenges we ran into

FLASK

Accomplishments that we're proud of

We were able to get flask somewhat working

What we learned

MongoDB, How to finetune an LLM, FLASK

What's next for Daily Journal

Our goal is to further develop and our project and add a chat bot to encourage more conversation. This could be helpful in a medical practice of trying to get large amounts of patients to talk about how they feel. This could also highlight patents that might need more attention or encourage an average individual to seek help if need be. We also want to connect our users with therapy in their area & help college students with their mental health.

Built With

• amazon-web-services
• distilbert
• flask
• huggyface
• kaggle
• mongodb
• pandas
• phi
• pytorch
• react
• scikit-learn
• transformers