We wanted an app that could tell us from an analytical perspective more about who we're talking to. As high schoolers, we use social media to communicator with out friends more often than not, and we began to see a common problem in our respective lives: they interactions in our lives have become increasingly boring. We wanted a new way of looking at how we talk to other people (and how other people respond), we wanted something more than just the text we see and send on our screens - so we created IBMessenger.
What it does
Our app can tell us the person's personality traits, their general sentiment towards and more. We also built in extra features such as healthy texting - if your general tone takes a sharp spike downwards or we pick up certain key words we can recognize if you're suffering from depression and get help.
How we built it
We used the Twillio API to message/communicate between the app and other people’s phones. All of the Twillio management is done on the backend with nodejs. After a sufficient amount of text is used in the interaction, we sent the text to our Watson backend which ran IBM Watson’s personality insights. We then use those personality insights to create a “personality profile” for everyone you message. Based on that info, we calculated a “compatibility” score which represented to other persons interests in you, along with other various information to provide a greater insight into your daily conversations.
Challenges we ran into
Connecting the Twillio to both receive and send SMS without being too complex. We also had some trouble setting up the initial IBM Watson backend.
What we learned
We learned a lot about the Twilio API as well as setting up IBM Watson.
What's next for The Ultimate Messager
We are looking forward to launching our app in the app store. Hopefully, with our app, people can look at the interactions in their lives in a new light!