An interesting thing to note is that with our data it that an increase in White residents doesn't correlate with any change in complaints-- that seems to suggest that the conversion are occurring somewhat evenly and have little to do with MHI changes. It would behoove us to perform a spatial autocorrelation analysis on areas with high Complaint counts in order to better understand what is driving them, and if it's geographically-specific (perhaps something to do with zoning or land-use?)

Conclusion

A first pass at the work indicates that further study, or a shift in methodology, is required. The data for shifts in ethnicity trend flat with shifts in complaints, so maybe we need to look instead at absolute complaints, and absolute numbers. A deeper dive into East New York and Sunset Park is suggested by the work above, since there seem to be significant pockets of illegal conversions occurring in both.

Future Work

There is plenty more work left to be done, including but not limited to: looking at more individual years, widening the year range to capture longer term trends, and refining the statistical analysis. I anticipate a thorough spatial autocorrelation study to be particularly fruitful in identifying dramatic outliers to general trends, that we can then focus on for further investigation. 

Links

Links to GitHub repository containing Jupyter Notebooks here: https://github.com/nlicalzi/PUI2018_nl949/tree/master/NYC-gentrification-and-illegal-construction