Inspiration
A lot of people do not seek therapy or share their problems because of lack of anonymity. Many a times, it also results in suicide. From college students stressing out on new adult life to old people missing their grandchildren - all need somebody to "talk it out" to. We enable this anonymously via TalkIt. Depressed humans get a virtual professional free of cost.
What it does
Its features include -
- Anonymous ChatRoom wherein our trained ChatBot takes over until a professional is available. The ChatBot is trained to handle emotions and talk like a friend and give out useful resources on analysing the person's mood.
- A Home Screen with daily tips and links to various articles
- ChatBot has Sentimental Analysis integrated along with Vision AI to analyse the user's sentiment and give a [beta] personalised wellness feed based on the sentiment score and better train the bot for future responses.
- [beta] Also uses Speech-to-text to analyse the person's emotions through their voice.
- Sends 3 daily notifications (firebase) with a positive/happy/funny quote or photo or video link.
- Displays a map of nearby mental healthcare centres
- Local resources like Suicide Hotline Number and other Organisations
- [beta] Use Facebook's Graph API Explorer to extract only funny content from Facebook based the maximum number of "HAHA" reactions on a post.
How I built it
We used React Native environment for android and various Google Cloud APIs and Facebook Graph API to make this project.
Challenges I ran into
We were having trouble setting up our environment until yesterday 8PM. I once set it up and created the ChatBot in the morning but it all crashed without any back up or way to go back. We spent endless hours trying to integrate the APIs smoothly - our major hurdle was the react-native-maps. 2 of our teammates backed out in the middle of the day with no code for us to use.
Accomplishments that I'm proud of
All of us have coded in React Native for the first time by mainly just going through the docs and relevant tutorials. We worked efficiently as a team by leveraging on our skillsets and dividing the work (features) meaningfully. We were able to partially cope up the work of other members that left.
What I learned
PATIENCE. This one word tells a whole story.
(And of course I learnt all the new technologies we implemented)
What's next for TalkIt - Mindfulness Manager
- Finalise all the beta versions
- Make the funny content personalised like the Instagram Algorithm. Example - one person shows more affinity towards dog videos while other towards prank videos.
- Integrate with Siri, Alexa, Google to give a "voiced friend".
- Integrate de-stressing games with a maximum time limit per day
Integrate Analytics to help user track their emotions and give a daily, weekly and monthly report.
Partner with colleges so their Health and Wellness Facility can be the professionals providing casual talk to students who do not seek help out of shame.
Scale it like a one-stop Mental Health NGO as big as Planned Parenthood.
FAQ
You said users get Virtual "Professionals" free of cost? Yes, our purpose is philanthropic. These certified professionals will either be volunteering, get donations, or be school centres that can utilise this as a platform for their students.
How did you train your ChatBot? We used the following datasets link as well as Google's pre-build Small Agent.
How did you get information of Health Care Providers? We used the data by New York Government for this version. link
Why is your app important? At some point, we all need some unofficial therapy. While there are many resources out there for people with mental health issues, our focus is not non-clinical mental health. We do not want to use our certified professionals to give a full-fledged free therapy, but to provide better communication and advice as compared to peers or nobody at all.
Why is your tagline "we're blindly listening"? Lol, that is to say that we're here for you and we're listening to what you want to talk, but we do it blindly (you're anonymous). We don't see you but we listen to you.
Built With
- ai-applied-sentiment-analysis
- cloud
- cloud-vision-ai
- data.gov.sg
- dialogflow
- express.js
- facebook-graph
- firebase
- google-maps
- graph-explorer
- healthdata.gov-catalog
- javascript
- jquery
- machine-learning
- natural-language-processing
- node.js
- npm
- python
- react-native
- socket.io
- speech-to-text
Log in or sign up for Devpost to join the conversation.