I was brought onto this project to provide insights on where and how a Real Estate Investment Company should invest within the United States. Using Zillow's open source time-series data, I am able to analyze every zipcode in the country to find out the best strategy to pursue.
The definition of "best" is going to be zipcodes that have the highest projected increases in values over the next 3 years.
94027 - $1,610,842.23 Projected Increase Home Values (24% ROI)
90210 - $1,485,517.35 Projected Increase Home Values (26% ROI)
94301 - $1,134,788.43 Projected Increase Home Values (25% ROI)
33480 - $1,107,069.31 Projected Increase Home Values (23% ROI)
94022 - $993,409.51 Projected Increase Home Values (24% ROI)
While I initially looked into creating my own SARIMA model with the data, I was not happy with the AIC and models ability to predict the home values in general. I looked into tuning the model until it became too computationally expensive and then opted to go with Facebooks Prophet Model to make the predictions shown in the Jupyter Notebook.