We are very interested in the idea of making computer understand human’s feeling from Mirum challenge. We apply this idea on calling center where customer support can’t see customers’ faces via phone calls or messages. Enabling the analysis of the emotional tone of consumers can help customer support understand their need and solve problems more efficient. Business can immediately see the detailed emotional state of the customers from voice or text messages.

What it does

The text from customers are colored based on their tone. Red stands for anger, white stands for joy.

How I built it

We utilize the iOS chat application from the Watson Developer Cloud Swift SDK to build this chat bot, and IBM Watson tone analyzer to examine the emotional tones, language tones, and social tones.

Challenges I ran into

At the beginning, we had trouble running the app on iPhone. We spent a lot of time on debugging and testing. We also spent a lot of time on designing the graph of the analysis results.

Accomplishments that I'm proud of

We are proud to show that our chat bot supports tone analysis and basic chatting.

What I learned

We have learned and explored a few IBM Watson APIs. We also learned a lot while trouble shooting and fixing bugs.

What's next for Chattitude

Our future plan for Chattitude is to color the text by sentence and make the interface more engaging. For the tone analysis result, we want to improve by presenting the real time animated analysis result as histogram.

Built With

Share this project: