Inspiration
As we got older, we saw more and more of our friends struggling with mental health and experiencing issues like depression and anxiety. Oftentimes, these people would not feel comfortable opening up with others, except for their closest friends, yet even being close friends, we often felt that we could not help them get through their difficulties. While therapy is usually the best often, it’s often stigmatized, awkward, and difficult to access, leaving people feeling alone in their struggles. With our app, we wanted to tear down the barriers to mental health care through organic connections between people and find support in an otherwise dark-feeling world.
What it does
Our app uplifts people by increasing availability to healthcare and strengthens support groups. When a user signs on, they choose the mental health groups they want to join, including Depression, Schizophrenia, and others. In these groups, they can post status updates and discuss things online, validating their journey to improving themself. Additionally, our app has an AI model that acts as a semi-therapist. This AI recommends proper steps that may help a person improve or feel better about themself and additionally recommends them to a psychologist if they display signs of serious mental illness.
How we built it
We split up our application into 2 separate components: a therapeutic chatbot and a community feature for making posts regarding mental health growth and challenges. The chatbot used GPT 3.5-turbo, and was urged to refer the user to a psychiatrist if they showed any signs of potential mental illness. The community post feature used repldb to split the users into various communities of mental health upon signup.
Challenges we ran into
While we were programming this app, we first of all struggled to implement our model in a manner that was both safe for the users and efficient. We had to figure things out as we went in order to create the AI, and we managed to make one that helps users. Additionally, we struggled with implementing our concept of people joining multiple groups and being able to switch between these groups. We had the option of making it so a person could only join 1 group, but we knew that many people have multiple mental health issues and restricting them to just one would limit their ability to get support. All in all, despite the challenges we faced we significantly advanced our app and made it into something we are proud of.
Accomplishments that we're proud of
Even though there is room for improvement, we are proud of the chatbot feature, as we struggled to make it function properly. Additionally, even though we didn't use a lot of time to work on the frontend it turned out looking quite nice, especially considering how we designed it for mobile phones. Overall, we are proud of the end creation.
What we learned
We gained a lot more experience working with the GPT 3.5 API to engage users in conversation. Additionally, we realized that our information sending to and from the backend was failing because it used the wrong method of URI encryption, which took many hours to fix and was ultimately enlightening. We also learned how to use ReplDB, which was very different from the databases we were accustomed to working with, which is primarily PostreSQL. It was somewhat different from MongoDB as well, but working with it was just different enough to need different code.
What's next for Mindful
In the future, we would like to advance our AI model so it is even more beneficial, and get survey responses from real testers to find the areas it lacks in so we can lead to a better overall experience. We would also like to improve the visual interface of the app so that users can more clearly navigate the process. Finally, we would like to add the ability for people to comment on posts made by others. This would seriously help users get aid and validation from the online community in a manner that is hindered by the current post-only system.
Log in or sign up for Devpost to join the conversation.