Breakpoint

• 

  Good text

• 

  Bad text

We build a custom keyboard that performs sentiment analysis on typed out text and pulls heart-rate data from a smart watch to determine how the user is feeling. It then changes the color of the keyboard based on the positivity/negativity of the user and also vibrates the watch when their heart-rate is also high. We built this to combat cyberbullying and improve mental health by warning the user if they're being negative or have anxiety. Both the visual indicator and physical vibrational feedback help in alerting the user of their emotions if they are unaware. We used Java in Android Studio to develop the keyboard application and javascript/css/html in Tizen to build the watch app. We faced challenges in extracting typed text as well as integrating IBM Watson's NaturalLanguageUnderstanding API. Programming for a watch for the first time, we also had trouble initially sending and receiving information from the watch. Through the entire project we learned that perseverance and grit pay off in building something truly impactful. Working with a team for our first hackathon, we figured out what it takes to work as a team and reach a realized goal.

This keyboard is based on AOSP LatineIME keyboard. You can get the original source code in https://android.googlesource.com/platform/packages/inputmethods/LatinIME/

Built With

• java
• javascript