PsychAid

The future is computational psychiatry. For psychiatrists to integrate AI into their practice.

Comment

Inspiration

I am a volunteer at a computational psychiatry lab at McGill. There is extensive research where researchers have collected data from patients using consumer grade wearables/Apple watch and used AI to predict how well they are or will be doing in the future.

What it does

The user can upload an excel file of a certain format and see predicted phq9 categories for depression (None-minimal, Mild, Moderate, Moderately severe, Severe) after 3 months. I've train a logistic regression model using PSYCHE-D dataset: https://zenodo.org/records/5085146

The user has to log in and upload an excel file with patient data and then using the model it predicts phq9 category after 3 months.

How we built it

I used Python (sci-kit learn and pandas for the AI part and Flask to integrate HTML with Python and create the login system

Challenges we ran into

I used Flask for the first time. I'm soloing it so it was hard! Also it took some time to find a dataset I could work with.

Accomplishments that we're proud of

Login system, Allow users to upload excel files of certain format to make predictions of phq9 category

What we learned

Learned to do hackathons solo

What's next for PsychAid

Better AI tools, SQL-based login system, promoting to doctors

Built With

Share this project:

Updates