Tracking the frequency and quality of my conversations during a typical week.

Dialog is a mobile app that records conversations (that last for more than 2–3 minutes) in a purely visual manner. After each conversation, I ask the person I’ve spoken with to pick a color and a shape that they identify with, and I rate the quality of the conversation myself. The Dialog app saves the date / time, selected shape, selected color, and my rating in a CSV file, which is later used to procedurally generate a short film that visualizes my conversations as colored shapes organized in 3D space.

In order to test the mobile app and output, I tracked my conversations for a week (March 5th — March 9th, 2018).

In order to preserve anonymity — while maintaining some level of visual consistency — I chose a set of fourteen colors and four shapes for people to use as abstract representations of themselves. I designed the third screen in a way that hid the fact that I was rating the conversation, and added a check that would wipe the previously recorded color and shape in case the data was invalid.

All four screens of the Dialog app.
Autogenerated data from the Dialog app.

While I tested the Dialog app during the week, I wrote a program that took the autogenerated CSV file (shown above) and translated each data point into a series of colored 3D objects that appeared and disappeared according to when the conversation happened. The final result was a 120-second video, shown below.

Each hour translates to one second in the short film, the elevation of the object is determined by the rating, and the horizontal position of the object is determined by the time of day. You can see the full piece below (featuring music and sound design by Yimin Zhang).

Final output from the data shown above!

If you want to download the Android app and try the demo yourself, you can find the code and assets here! As always, feedback and questions are welcome: sagarram [at] usc [dot] edu.