Identify areas of need in the community and visualize a big-picture view of how the organization might address the identified need.
Identify possible cause(s) of the overarching problem so that we can begin addressing the need.
Determine what needs to change and how, addressing the identified cause(s) of the problem.
Connect organizational activities to intended change, and identify data sources for measurement.
We wondered whether there was a way to predict eviction rates in Tulsa County. The first step was to understand the variables that influence eviction. Using machine learning to analyze eviction data, we identified 36 variables that are common distinctions between individuals who face certain levels of eviction risk. Among those were variables that we did not expect and would not have included if we were coming up with them ourselves. Identifying variables using machine learning is a necessary first step in understanding eviction risk so that we might be able to pinpoint the cause of evictions in future research.
Start a conversation