Projects‎ > ‎


Zillow is a real-estate information site which provides APIs to extract real estate, home value, profile and mortgage data.  For any 'neighborhood', can you predict which quartile a house price is based on property and local area features? Which house prices do you predict to belong to quartiles above their observed quartiles?

1) Data Extraction:
1.1) Review the instructions for accessing the Zillow APIs
1.2) In a browser, first experiment with using the different APIs using the API key X1-ZWz1dexzianjm3_9od0r
1.3) Use the urllib2 module in python to extract the relevant content from the site based on your definition of a 'neighborhood'.

2) Feature selection and model building:
2.1) Build a supervised learning classifier to decide which quartile a house price belongs to.
2.2) What are the most important features?
2.3) What properties of the data motivate your choice of classifier? 

3) Classifier evaluation:
3.1) How well does your classifier perform out-of-sample?
3.2) Repeat above steps across different 'neighborhoods'.