With many Canadians facing mental health issues, we wanted to create a daily journaling app, which uses machine learning to recommend resources for mental well being.
What it does
After a user records a 10 second video talking about their day, Menda uses emotion detection and facial recognition to recommend resources. Users can see a daily log of their well being, in which Menda curates personalized suggestions over time.
How we built it
We used HTML/CSS for the Frontend, Firebase and Flask for the Backend, and OpenCV/Nltk for machine learning.
Challenges we ran into
The biggest challenges that we faced were connecting ML models to Flask, and building entire application around Flask.
Accomplishments that we're proud of
We are proud of our emotion detection function, sentiment analysis from speech to text, as well our minimalistic UX design.
What we learned
We all took advantages of opportunities to improve on our technical skills. Although we've all participated in hackathons before, we still each picked up new skills. For example, everyone working on backend had the opportunity to experiment with Flask and Firebase, while those on frontend were able to enhance their HTML/CSS/JS skills.
What's next for Menda
First, we want to create a community page, where others can share resources and discuss among peers. We also want to consider what happens as we increase the number of users since we store videos within our database. So, in order to scale, we would have to upgrade our Firebase database to be able to store more data. Lastly, we want to look for partnerships with related mental health organizations as well as apps that we utilize in Menda such as Spotify and Headspace!