Inspiration
Messaging apps are an important medium to connecting with friends, family, and loved ones. Whether its for keeping in touch with a loved one overseas, or simply roasting your buddies, the communication that these chat apps buttress is a real tangible thing.
Our team recognized an opportunity to create more value out of the messaging app platform, by using a machine learning algorithm to contribute added value to messaging. Using the same algorithm that can decipher if an email is spam or normal (Naive Bayes), we show that a simple model can recognize text message emotions and feelings, and the app itself can change as the user inputs novel messages!
What it does
This project is a demo to be run locally on your computer. Message a different 'user', and the message itself will be classified based on what our machine learning model predicts the emotion of the text to be. The text color will change based off this underlying tone of the message.
How we built it
HTML, CSS, and Javascript were used to create a simple web page that creates a chatroom and can listen for local users to input text data.
We coded an implementation of Naive Bayes using Javascript and trained emotive text messages with the algorithm
Challenges we ran into
We started very ambitious and wanted to add full server-side functionality, but these proved to be too much to tackle for our team. We originally wanted to have our app communicate with remote users, but we changed our goals to be able to sign in to our webapp and show the emotions of texts. If we had more time in the competition, we would include more php and MySQL functionality into the application (but this project will be improved outside of the competition)
Accomplishments that we're proud of
For our group, this is the first somewhat functional web app that we have EVER made! Merging the front end and the back end was a huge celebratory moment for our group!
What we learned
We definitely have encountered a greater appreciation for the skill needed to pull complex webapps together from scratch. Even with the time we had for this hackathon, there are still countless ways that this application can be made much better, from improved aesthetics to server side functionality.
What's next for Machine Learning Messaging App
We want to continue working on making this concept better for the next upcoming year! We plan to apply to industry internships, and this project gives us an opportunity to learn cool technologies that will be important for those responsibilities
Log in or sign up for Devpost to join the conversation.