positiviTree

This project aims to improve the mental health of users, using positive and self-affirmation. Research indicates that the positive affirmation of users ‘good’ thoughts and opinions can lead to a more positive mental outlook. This decreases their stress, increases well being, improves academic performance, and can improve openess to behavioural change.

Concept

• Positive reinforcememnt of ‘good’ thoughts
  1. The user speaks their mind to the web-app.
     • Speech processed by natural language processing.
     • Sentiment analysis is processed.
  2. The positiviTree grows or shrinks depending on whether the thoughts are ‘good’ or ‘bad’.
     • Additionally, phrases which are strongly indicative of poor mental health trigger a warning, and respond with a helpline.
  3. The feedback from the positiviTree positively affirms the user’s ‘good’ thoughts, and gives negative feedback for bad self-image.
• Responses aimed at helping the user talk
  1. UNK

This project was bootstrapped with Create React App.

Init

• npm install
• node node/server.js
• npm start

DevPost Summary

Inspiration

Mental health is becoming more and more of a recognised issue in the current media climate. Social media particularly has played a complex role in affecting individual’s self-esteem. Multiple studies have linked high social media use to people with social limitations and insecurities. We hope that our web-app provides a lightweight and friendly method for helping people deal with issues around self-esteem and social anxiety. By providing simple interactive app that encourages optimism and self-confidence, we hope to encourage positive behavioural changes and help stabilise vulnerable persons.

What it does

The web-app prompts and listens to the user speak about anything. The app rewards comments it analyses as positive, by growing the adorable positiviTree and giving it a smiley-face. For negative comments, the app shrinks the positiviTree and gives it a sad face. When the app detects highl concerning sentiments from the user, a helpline to samaritans is provided and the user is urged to contact them.

How we built it

• The frontend was developed using a reactJS framework.
• Graphics were designed by Sarah, and animated within the reactJS framework.
• Audio is recorded by the website, and piped to the google API for transcription.
• The raw text is then piped to a backend server hosted on nodeJS.
• The nodeJS server launches a child_process in python, which runs sentiment analysis using the vaderSentiment package.
• Sentiment analysis is returned to the frontend, along with the warningLevel (in case the user is in urgent need of mental health support).
• These values then instruct the ‘growth’ or ‘shrinkage’ of the positiviTree.

Challenges we ran into

• Implementing the ‘growing’ and ‘shrinking’ of the positiviTree in react proved tricky.
• Capturing and successfully piping audio from the frontend to the google API also proved an obstacle we had to work around.

Things we learned

• ReactJS
• Manipulating microphone audio on the frontend using AudioContext
• Versioning with git

What’s next?

• Deploying the web-app to a dedicated server.
• Deploying the backend to a dedicated server.
• Developing a user database, and more personalised prompts to encourage positive thinking.

Available Scripts

In the project directory, you can run:

npm start

Runs the app in the development mode.

Open http://localhost:3000 to view it in the browser.

The page will reload if you make edits.

You will also see any lint errors in the console.

npm test

Launches the test runner in the interactive watch mode.

See the section about running tests for more information.