Zillow Home Value Predictions

Project Brief

This project aims to help Zillow better predict its housing market predictions through a hedonic model. The project emphasizes on local knowledge, a novel way of looking at the data and creative factors that might enhance predictability of Zillow’s modeling. While Zillow’s model works satisfactorily, this approach offers a different lens into how the model could be built stronger with intel from a variety of internal and external factors, such as amenities like schools, parks, public spaces, as well as demographic data like poverty rates, median income, crime, etc.

In addition to the hedonic model, this project focussed on cross validation through evaluating mean absolute error/percent error and conducting Moran’s I testing. The full analysis and model can be found below or by clicking the Full Analysis link above.

Exploratory Analysis

The project conducted a very thorough exploratory analysis with feature engineering on a range of variables. Some external variables are visualized below:

Play streets are visibly more concentrated in center city, with other nodes of concentration across the city.

Grocery stores are well-spread throughout, with certain pockets of concentration in the typically dense areas (center city, university city) as well as a particular node in North Philadelphia.

Crime (the crime identified in this set are those self-chosen as crimes that would most affect residential areas, i.e. violent or property crime) has a distinct concentration in South Philadelphia and parts of North Philadelphia.

The greatest density of trees exist across central Philadelphia, across center city and university city, as well as various parts of northwest Philadelphia. Trees are more scarce in other areas.

Our final model resulted in a prediction home value with this level of similarity.

Project Outcome

Click the full analysis button at the top to view at full screen.

Data Sources

• Zillow, wrangled by Michael Fichman here
• OpenDataPhilly
• Census Bureau