A mental wellness app that allows users to "plant" thoughts and develop them positively.
Our team views mental illness as a collective issue: one that is intergenerational and is exacerbated by conditions of poverty and racial injustice. As a result, mental healthcare is not adequate or affordable for many. Our app is part of an emerging realm of e-mental health applications, which is an initiative to act as a stopgap for comprehensive mental healthcare, deployed in a way that is more accessible to all. Wellspring is a mental wellness web app that allows users to confront their pressing thoughts in a healthy, guided manner. Users are prompted to enter what's on their mind, followed by a short description of how typing that thought made them feel. The users then asociate a color with this emotion, and are given the space to free-write. After free-writing is completed, the user "plants" the thought, all text fades, and flowers bloom. The text is simultaneously used to fortify a sentiment analysis model. The floral arrangement and pattern is determined by a combination of the user's reported and predicted sentiments.
How We Built It
The frontend is built with Flask. The text input and user data is stored on MongoDB. Our app uses a logistic regression model for sentiment analysis built with scikit-learn. This model was then uploaded as a model server to Spell.
What We Learned
Publicly available population mental health data that can be obtained without research credentials is surprisingly hard to find. Machine Learning is tricky! CSS is powerful.
How We Would Develop Further
Finalizing the integration of our data stored in MongoDB Stitch with our model in Spell. More functions used to manipulate and draw conclusions from the data stored in MongoDB. Use of larger, more detailed datasets (such as those from crisis text lines). Further aesthetic development of the frontend.