What scales of narrative lie in data about a city? Humans of Simulated New York is a participative & speculative economic simulation and game based on 10 years worth of New York Census data. You start the game being assigned a “character” — a New York citizen, and you vote in a group of three on what legislations to propose, so that you can raise your quality of life. We basically made a simulation of SimCity, in New York, and replicated its inhabitants.
Experience the simulation and tweak it's knobs here. Project done with Francis Tseng.
HOSNY started like a data visualization. We had our NY census data and information about NY citizens’ employment status, industries of employment, salaries, age, race, gender, and education levels.
People are color-coded according to Census-coded race, used Flatland’s hierarchy of shapes (polygon count correlated with economic status). Building layers are color-coded according to the businesses that occupy them (one layer per business).
They became simulations when we used those details to generate another set of “plausible” New Yorkers with machine learning. We give these “agents” very simple “personhood”. They are just as likely to get hired/fired as their real life counterparts because their generated selves are based on real labor statistics. Their role in socioeconomic dynamics are preserved.
The sim-New Yorkers live inside a city with a basic economic system where the industries are “agents” and so is the government. These agents use what’s called Q-learning, a type reinforcement machine learning to modulate themselves.