Shouldr

A mental health chatbot to talk to about anything

Comment

Inspiration

The inspiration for this project came from the Hackathon's theme of social good. This year has proved tough and we wanted to create something that could get everyone opening up and point people to the right help if it seems like they could use some professional help.

Chatbots are increasingly used globally, and there do exist therapy bots. However, these therapy bots can be very prescriptive, asking the user to complete exercises. While this is one form of CBT, there is an alternative of talking therapy, with more listening and less goal setting. For the current COVID situation, we thought it would be nice to implement a talking therapy bot where people can come to open up and think about their own mental state. In this way, Shouldr is encouraging development of emotional intelligence by leading the user to go into detail about their problems.

What it does

Shouldr is a chatbot to help users start opening up their minds to their feelings, identify mental health issues and monitor wellbeing. It is in no way trying to replace traditional therapy; rather, Shouldr provides a shoulder to lean on, to merely listen with no judgement. The aim of Shouldr is to use NLP to understand the feelings of the user and to then help them explore why they feel this way.

How we built it

We split into two sub teams, with Deveena, Theo and Sorana working on the backend and Jeff and Krisha working on the front end. To implement the natural language processing, we used IBM Watson's API to perform sentiment analysis on the user's input, and then based on this result we use a database to select the next message from Shouldr to the user.

Challenges we ran into

The main challenge we have faced is connecting the front and back ends due to time constraint. We did not already have the skill in the group so to do this we would have to develop this skill with very little time. We have decided to improve what we have so far and would look into integration of the ends over a longer time span, or find someone with the relevant skills to take a look and advise on how this could be achieved.

Accomplishments that we're proud of

Getting our chatbot up and running and our front end design.

What we learned

On the front-end, we learnt and used Vue js framework for creating application components.

What's next for Shouldr

Due to our time constraint, we have only been able to create what we feel is merely the starting point. There are many directions that can be taken with the current framework. Here is a non-exhaustive list for next steps

  1. Expanding the database
  2. Using more analytic features provided by Watson, including emotion analysis, their personality framework and categories and concepts
  3. Using location to find relevant mental health services near the user
  4. Categorise text, e.g. if suicide is detected, then direct the user to a suicide hotline
  5. Based on the user's personality traits, tailor responses
  6. Store chat histories

Built With

Share this project:

Updates