Everyday we hear about a new poll, we read a new study, we’re told about new statistics that show how every piece of data that describes us alters our future: how the food we eat, the cigarettes we smoke, where we live, our gender, our age, our car make is going to alter predictions about our future life, the money we earn, the time of our death.

What we wanted to do is show people what changes what and by what amount. We wanted to give users the opportunity to see how their data defines who they are for systems. We wanted to let people be the system in a way, and ultimately play with it.

Predictive World is an online experience that reveals what a powerful, all-seeing algorithm can predict about an individual based on their online footprint, public information and Facebook ‘Likes’.

The dangers of Big Data and predictive algorithms are themes at the heart of the narrative of Ubisoft’s upcoming open world action-adventure game Watch Dogs 2. The hero, Marcus Holloway, is wrongly profiled as a potential criminal by a city-wide operating system collecting & analysing data on every citizen. Therefore, it was clear from the beginning that data would be at the core of our creative.

Harvesting the data was our main concern, so we partnered with the Psychometrics Centre of the University of Cambridge to help collect, gather and clean the necessary information.

Initial predictions are made on the basis of age, gender and location by a series of calls to large databases (WHO, Unesco, OECD, etc.) that were constructed specifically for Predictive World.

Psychological predictions are generated via a set of proprietary algorithms from the Psychometrics Centre that were calibrated for the experience.

One of the main challenges was to display a massive amount of data while addressing a mainstream, non-educated audience. On desktop, we went for a very graphic and abstract eye-shaped interface, reminiscent of a cyber-system eye that would see through the data of the user. A console at the bottom right of the screen constantly explains to users the data shown on screen and encourages them to interact with it.

On mobile, we opted for a more streamlined and straightforward interface that would drive the user on a simpler narrative through a number of successive screens.

Users could play and modify their data to see how the subsequent changes would reflect on their perception: measure the impact of heavy drinking on life expectancy, their aggressiveness on the chances of getting the job, and discover unexpected correlations such as height and annual income.

After 8 months of development, Predictive World was launched on November 10, 2016.