Understanding Empathy - Julie



Being a technology enthusiast, I tend to be an early adopter, experimenting with technologies that improve our daily lives. Though personal assistants like Amazon Echo, Google Home and Siri are increasingly gaining popularity with them getting better at executing instructions like switching on/off our lights, I can't help but notice that there is one alarming common issue that they have not addressed -- understanding empathy.

This thought is also highlighted in a recent CNET article released late last year, indicating that when researchers from Stanford University and University of California made health statements such as "I am having a heart attack" or "I want to commit suicide" to digital assistants, they found out these digital assistants only responded with variations of "I don't understand".

Given the potential pervasiveness of personal assistants, I believe that this is a problem area that is worth exploring deeper on how to address.

As part of a mobile IoT class, our final deliverable was to create a prototype mobile app that links up with smart devices that solves a real consumer problem. As such, my team of 8 got our hands dirty, and created a prototype of our product, Julie, within 3 weeks.

We named our app Julie because the name gives the feeling of a friend, someone close that can lend a listening ear. Also, you would have noticed that we gained design inspirations from the characters of Inside Out, a Pixar film that beautifully explores inner emotions and its reflections at our outside world.

In a nutshell, our prototype iOS and Android app allows users to enter their emotions through three methods: emoticon icons, voice input (that we used APIs to convert to text and through text get sentiment analysis), and facial expression (we used google vision API to get the sentiment). After that, we linked the emotions to a library of appropriate songs to play, and a Philips Hue custom API that triggers the light bulb to change intensity accordingly.

My main role in this project was to be the product manager looking for market fit and managing task allocations with Scrumban methodology using Trello as our tool to manage the product backlogs in weekly sprints. Also, I coded the APIs of the Philips Hue lights for the mobile apps to connect to.

Details of our story board, project management, and architectural diagrams can be found in the slides below. I have also attached two YouTube demo videos we recorded at the back of the slides for your reference. Enjoy!



Overall, the prototype received good feedback from the class. We think that there is potential to improve this product further, i.e. linking up emotions to recommended activities like favourite food places, hobbies, groups of friends, or healthcare providers to address emotions. Also, with the data collected, we could do some trend analysis to improve song recommendations. However we did not continue to pursue it as the team had other commitments after the class. 

Feel free to comment about this product at the comment section. Cheers!



No comments

Powered by Blogger.