There are a multitude of pressures permeating the current climate of fear in our society, including loss, unemployment and isolation. All of these problems and tragedies undoubtedly have an effect on our mental wellbeing; in fact, a recent survey of 800 people in the UK living with mental illness found that 80% felt that their mental health was much worse due to the pandemic’s impact. Furthermore, mental illnesses often go undiagnosed and untreated, resulting in an emotional and economic cost. The third Sustainable Development Goal promotes good health and well-being for all, and we aim to propagate this mission with our app that addresses the question, ‘How might we use AI to improve the mental health of people aged 14 onwards, during the pandemic and beyond?’
What it does
MoodSmart is a cross-platform mobile application that aims to alleviate cases of mental health problems. This is done by analysing a user’s social media posts and detecting early indicators of negative mental health. In particular, we focused on depression for our prototype as it’s an extremely pertinent illness affecting 264 million people worldwide.
How we built it
The back-end, a Python Flask API, uses a Machine Learning algorithm that employs Natural Language Processing, in order to perform sentiment analysis on tweets. The front-end was created with Flutter.
Challenges we ran into
We found it difficult to query the Flask framework through Flutter but we successfully accomplished this. In addition, communication between all team members was hard to manage, but we successfully got around this.
Accomplishments that we are proud of
We are proud of our appealing UI design and a machine learning model with 99% accuracy. Moreover, it was greatly insightful to research into mental health and sentiment analysis.
What we learned
(Demi) I learnt how to be resilient when things don't go as planned. I also learnt how to be a great leader and allocate roles accordingly (Grace) I learnt some many new techniques using Python, Flask and Flutter, even completing a Coursera course! Additionally, I learnt how to manage my time effectively to produce an effective prototype. (Dinali) I learnt how to research effectively, find statistics and data relevant to our topic of mental health and depression and develop part of a concise script for a pitch. (Criofan) I learnt how to create a distinct logo and effective video for a pitch.
What's next for MoodSmart
Future plans for MoodSmart include integrating it with other social media apps such as WhatsApp, Instagram and Facebook. We also want to include online therapist sessions for users on a free trial basis to assist the users during these hard times. Furthermore, we would like to improve our model to detect other conditions such as anxiety.