Many lack access to mental health resources due to prohibitive costs and lack of access. We seek to mitigate this problem using technology.

What it does

Each day, users reflect verbally to the application. The app transcribes their reflection and uses sentiment analysis and Part of Speech detection to parse the user's input and understand their mood and any root causes of their negative emotions. Based on these identified root causes, the app prompts the user with follow-up questions to better monitor their mental health and guide the user just as a real therapist would.

How we built it

We created a flask web app that used the nltk library to perform Part of Speech tagging, and used the IBM Watson API to perform sentiment analysis. The mobile app was built in Android Studio and includes a fluid user interface. Firebase's Firestore software was used to store users' reflections.

Challenges we ran into

It was difficult to successfully perform part of speech tagging due to the myriad of ways any given word can be used, and due to the irreliability of the APIs. It was difficult to integrate the Flask webapp with the android project as well.

Accomplishments that we're proud of

We firmly believe this idea could have a positive impact on many people. We successfully perform sentiment analysis on users' reflections and reliably track analytics regarding their progress over time. We were able to incorporate a proof-of-concept of a more conversational, virtual-therapeutic experience.

What we learned

We learned how to create Flask web apps and use the nltk library as well as IBM Watson to perform difficult natural language understanding tasks. We learned more about their underlying algorithms and further developed our skills at crafting smooth user experiences.

What's next for Captain's Log

We hope to keep working on our project and refine it such that it provides a very interactive, therapeutic experience for users. We'd also like to involve mental health professionals in analyzing the data collected by our software. We also believe the data we collect can be useful for mental health diagnosis.

Share this project: