Purpose
TypeSense is a chrome extension that analyzes and displays a Facebook Messenger conversation's sentiment in real-time, designed to help the developmentally disabled. Winner of the best healthcare-related OSS hack at HackIllinois 2018.
Installation
TypeSense is available for download on the Chrome Web Store. Once installed, TypeSense requires no further setup – just open a conversation and click the TypeSense icon to visualize its sentiment.
Why?
Conversations can be ambiguous. They're even more ambiguous online, where the body language and intonation cues that usually help us interpret meaning are lost. People without neurological disorders misunderstand each other all the time – sarcasm makes it hard to know for sure what someone means, and idioms are confusing! Imagine how hard it must be for people with social disabilities.
TypeSense makes it a little easier for those with developmental disorders, like autism, to understand the emotions of others when chatting online.
Features
- Visualize the conversation's sentiment over time
- See how each message impacted the overall sentiment
- Real-time updating of the TypeSense icon's color based on the sentiment of the most recent message
Contributing
To make a contribution, fork the repo, make your changes and then submit a pull request. If you've discovered a bug or have a feature request, create an issue and we'll take care of it! :)
Pending Features:
- Analyze and visualize the emotional tones of the conversation (anger, disgust, fear, joy, and sadness)
- Analyze and visualize the language tones of who you're talking to (analytical, confident, and tentative)
- Analyze and visualize the social tones of who you're talking to (openness, conscientiousness, extraversion, agreeableness, and emotional range)
- Predict the impact of messages on the conversation's sentiment before sending them
Technologies
TypeSense is built on VADER, a lexicon and rule-based sentiment analysis tool that is specifically attuned to sentiments expressed in social media. We plan to release some more features, noted above in pending features, that will leverage IBM Watson's natural language API.
Team
Developed and maintained by Jonathan Shobrook and Aaron Lichtman.
Prototyped at HackIllinois by:
Donations
This is free, open-source software. If you'd like to support the development of future projects, or say thanks for this one, you can donate Bitcoin at 1BFqZayDhiU9YjhTuAsWcnu6qU5RDaxV8Z. Everything is appreciated!