Long Beach Innovates

Tagreted Strategies for Reducing Homelessness

Cellular automata divides a larger area, such as Los Angeles county, into smaller areas, such as census districts or voting tracts, and uses rules to propagate what happens in one cell to the next.
High crime or homeless in one cell? How do we propagate that to the nearby cell? Is the propagation only based on distance or is it also based upon demographics?
This effort attempts to incorporate census demographics and nearby crime or homeless levels to inform and/or predict where the near-future high crimes or homeless are most like to be. This allows politicians to form policy and law enforcement to more efficiently deploy resources and potentially head-off crime.

Time to party

After a Cerebral Beach hackathon, there is nothing to do but go party.

At the hackathon we improved our cellular automata simulation of the homeless population by census tracts.

This gives policy makers the chance to try different scenarios at solving the homeless propblem. Do you go after the census tract with the highest number of homeless people? Do you try to surround the highest census tracts from the perimeter and work inwards? Or other strategies.

This map shows the highest rate of homeless people per census tract for Los Angeles county. By proceeding with time with simple rules derived from neighboring census tracts, the homeless population in each census tract can change. Different rules can be used to try and correlate the simulation/model with reality.

Cellular Automata Simulation

This map shows the highest rate of homeless people per census tract for Los Angeles county. By proceeding with time with simple rules derived from neighboring census tracts, the homeless population in each census tract can change. Different rules can be used to try and correlate the simulation/model with reality.

Cellular Automata Simulation

We'll Try Again!

We wanted to wrap some artificial intelligence around the cellular automata simulation to give us a chance to look at various policy scenarios. Unfortunately the Kindo AI, as provided by one of the main sponsors of the Cerebral Beach hackathon, did not play well with Python. We'll try again.