A tone analyzer app submitted to the YHack hackathon at Yale University

Client: Mirum Agency, YHack Hackathon

Timeframe: Dec 2017 (2 days)

Role: Product design


Two non-coders went to a coding competition at an Ivy League school…

My first major hackathon was YHack at Yale University, where I teamed up with another designer. Responding to a challenge brief from Mirum Agency, we designed Emotionyze – a mobile app that analyzes the tone of a conversation and suggests a course of action. I designed the product and the UI, and created an interactive prototype. Emotionyze advanced to the second round of judging.

‣ Discover


My partner and I decided to talk to the Chief Marketing Officer of Mirum Agency, John Baker, who was at the hackathon. We discussed his organization, and asked what areas of opportunity he saw for tone analysis.

From speaking with John, we learned that an important touchpoint for brands is customer service representatives. To deliver great customer experiences, customer service reps need to know how to respond to tricky customers over the phone and on social media. Without the aid of nonverbal cues, it can be difficult to decipher a customer’s emotions.


My partner and I chose Delta Airlines to focus on for our research. We gathered information from customer reviews on Facebook and Trip Advisor, and kept these research questions in mind:

  1. What kinds of experiences do customers have with Delta Airlines?
  2. What sentiments are expressed towards the company?
  3. What touchpoints and interactions with Delta employees are mentioned?

Finding common themes, we created two quick snapshots of a positive and negative customer experience.

‣ Define


Customer service representatives


Understand a customer’s emotions and get suggestions on how to best respond


Help brands deliver great customer experiences


Which channel of communication to focus on — phone calls or social media? We decided that phone calls had a greater need for our app because there is the need to respond right away. With social media, you have the chance to re-read the post and plan a good response.

Emotionyze will use the IBM Watson Tone Analyzer API, which detects communication tones across three categories:

  1. Emotional Tones: joy, anger, sadness, disgust, fear
  2. Language Tones: analytical, confident, and tentative
  3. Social Tones: openness, conscientiousness, extraversion, agreeableness, emotional range

‣ Design


I sketched out some ideas to visualize the different tones. Each idea had pros and cons.


I designed the high-fidelity screens in Sketch.


Click through the prototype to try Emotionyze yourself!


The developers who were supposed to be on my team were busy with school assignments, so the team became just myself and my designer friend. Instead of giving up, we took this project as far as we could, and presented our InVision prototype to the judges. I demoed the prototype on my phone, and it looked so much like a real app that the judges even asked how we coded it! We advanced to the second round of judging, and I’m proud of how my first major hackathon went. Emotionyze was designed for customer service representatives, but I see future applications for social workers, therapists, and police officers.

“It’s refreshing to see a hack with so much attention to good design.”

- John Baker, CMO of Mirum Agency