Note: this project was a runner-up for the DocuSign Security Prize at CalHacks 2019
Our team thinks it's awesome that mental health is slowly becoming more valued in society, but a lot of the solutions that we see in the market today just aren't targeting the problem in the right way, considering the might of the 21st century technology we have at our disposal.
MindMap believes that the Psychotherapist has always been, and will continue to be, the key to improving mental health for most individuals. That's why MindMap gives therapists access to state-of-the-art data analytics tools to be able to best suit the needs of patients.
What it does
When a patient logs into MindMap, they're prompted to sign in to their Google and Facebook accounts, then asked to Docusign a waiver presenting temporary data analysis rights to MindMap and their therapist. MindMap then fetches their browser and messenger history on which 2 layers of NLP sentiment analysis are performed, quantified, and stored in a database.
Now, the therapist can use their side of the web app to look at high level data visuals as well as the fine-details and NLP results. They can create notes and discuss long-term emotional trends with their patient. By mapping their patient's mind with modern data analytics and NLP, they become a better therapist as a whole.
How we built it
We have two layers of NLP, one part through GCP's Sentiment Analysis API and another custom NLTK model written for further emotional analysis. Our database is SQLite, backend in Flask, and frontend in React.
One issue that arises when data like Messages and search history is used for analytics is the problem of privacy. We wanted to maintain the patient's privacy while ensuring that their physician can still provide them the best care possible. That is why we drafted a Therapist - Patient agreement in DocuSign where the patient can easily understand and consent to the forms of data and analytics that their doctor will have access to. This also amplifies patient ease-of-use since they just are sent the proper contract by their physician via DocuSign and they can understand it and sign it, and their data becomes available for the therapist to use for treatment.
Challenges we ran into
We pivoted 3 times during this hackathon due to API and software incompatibility issues. There were also major text parsing errors with the text data during the last few hours.
Accomplishments that we're proud of
However, we're very proud of what we accomplished in the last 9-10 hours of the hackathon.
What we learned
We gained a deeper understanding of NLP, encoding, and the importance of deciding on an idea beforehand.
What's next for MindMap
We're discussing how MindMap, after some bug fixes and feature enhancements, could become a market-ready SaaS for real psychotherapists.