Inspiration
Anxiety is a common mental health issue that affects approximately 63% of college students in the United States (according to the American College Health Association Fall 2018 National College Health Assessment). From our personal experiences, we know that music can be a refuge for people to deal with stressors in their lives and wanted to provide a more personalized and accessible method for addressing this issue.
What it does
The application prompts users with a series of questions from surveys commonly administered to determine anxiety levels in participants; it also takes in information from a linked Fitbit and looks for deviations from regular patterns in heart rate. Once a change is detected or the user indicates a certain level of anxiety, the application will recommend a song for the user to listen to.
How I built it
The application was built using React Native (JavaScript and CSS).
Challenges I ran into
Initially, we planned to utilize APIs such as the Spotify API or the Fitbit API to provide real-time input and output for users; however, it was difficult to find support for these capabilities, especially in the Python wrapper that our team was attempting to code with. We also had the idea of using a machine learning model to train our recommendation system, but ran out of time to execute.
Accomplishments that I'm proud of
Establishing the general infrastructure of the application for future iteration was our biggest success on this project; as long as we have the two-part detection/solution model, we can continue to modify and improve upon it even if the application is still unpolished for now.
What I learned
A key lesson was the amount of time it takes to iterate on phases of applications like these; tasks that we thought would take only a few hours ended up taking the entire night.
What's next for Sunshine
There are so many opportunities to iterate on the fundamental detection/solution model: implementing additional solutions to cater to a wider audience (article recommendations, simple notifications/buzzes to calm users, puzzles or games to distract a user from anxiety-inducing factors) or creating a more comprehensive input model to better assess when a user is anxious or distressed (real-time measurements, overall physical activity, levels of social interaction with friends).
Log in or sign up for Devpost to join the conversation.