• To provide an easy way to understand the mood and sentiment of a person based on typed messages. A common issue with text messaging is that the tone of a message can be difficult to interpret without the natural and inherent conveniences of spoken language. Using the IBM Watson Tone Analyzer, the goal of Sherlock Message Service is to bridge the gap by providing a simple, inline addition to the common text messaging UI

What it does

  • Using IBM Watson's Tone Analyzer, SMS returns and displays the tone of a typed message to convey intended sentiment - be it joy, anger, sadness, disgust, or fear. In addition, using Watson's Personality Insights API these messages are used to generate personal and private insights based on the Big 5 Personality Traits - Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism.

How we built it

  • For the front-end, our team leveraged React to create a prototype version of the Sherlock Message Service UI. The front-end handles the send and receive of typed messages, service calls to Watson APIs and dotNet, and post service to a SQL database. SQL is used as a means to store messages which are used to generate personality insights data. In a future version, our goal would be to remove stored messages to ensure the security and privacy of the user.

Challenges we ran into

  • We had a hard time implementing Watson text-to-speech and resolving a conflict between personality insight and the database. Also, the minimum 100-word count is difficult to work with inside a text messaging context.

Accomplishments that we're proud of

  • Combining Watson's data with typed messages to display message tone in a clear and simple way.

What we learned

-We learned a lot about the Watson Tone Analysis and Personality Insights API. Primarily how to receive response data, append it to individual messages, and display in graph format.

What's next for Sherlock Message Service

  • Ideally we would like to see our service provide tone analysis in real-time, prior to submitting of the message. This feature would provide a user the opportunity to adjust a message if desired. Further ambition would be to include Watson text-to-speech to provide additional accessibility features.

Built With

Share this project: