Study-on-Factors-influencing-NYC-Housing-Prices

Enigma Datathon Runner-up

• Extracted and Processed five NYC open data to characterize a location in NYC in term of approachability, safety, standard of living, etc.
• Combined these location relevant features with the housing price dataset to enhance model's performance and understand factors driving the house prices in NYC.

5 different publicly available datasets were used to analyse the factors influencing the housing price.

Datasets are :

1. Enigma NYC Property Sale: Accessed value land, Floor area residential, Floor area total building, Year built, Residential Units, altering history Gives information about the particular house and its properties.

2. FourSqure Data: Gives following venue information about Zipcode: Professional places, shops&services, Residence, Outdoor&Recreation, College & University, Travel&Transport, Night life&Sports, Arts & Entertainment. Gives idea about how engaging one area is

3. USZipcode API: Converts latitude and longitude information to zip code Gives other factors like : Number of House Units, Density, Land Area, Water Area, Population, Total wage, Wealthy Gives intuition about quality of life and population and demographic information

4. NYU Yellow Taxi: Number of pickups and drop offs from particular zip code Gives information about how accessible area is.

5. NYC Crime Data Gives following information based on Zipcodes: Felony, Misbehavior, Violation, Misdemeanor Gives intuition about safety of an area

Correlation map of all variables of final Dataset:

Feature importance:

• Gross square feet (Enigma Dataset)
• Commercial Units(Enigma Dataset)
• Density(ZIP Code API)
• Art & Entertainment (Fousquare Dataset)
• Police precinct(Enigma Dataset)
• Felony (Crime dataset)
• Outdoor & Recreation (Foursquare Dataset)
• Land Square feet (Enigma dataset)

Insights From Feature Importance:

• As expected, diverse datasets did give some idea about the different attributes that correlates with the price of a house, which we could also interpret as making a place for desirable to live.
• Intuitively Gross Square Feet and Commercial Units are important trivially.
• A high density naturally suggest high demand for living in the area.
• Felony and police precinct can be used as a proxy for how safe the neighbourhood.

Conclusion:

We focussed our efforts on finding the attributes that primarily drive the price of a property. As as natural next step, we would like to gather more data and built a model to predict prices primarily based on the features which we found to be important.