Objectives:

           The objective is to analyse the spatial distribution of  real estate price and its variability over time. 

           Real Estate is a special type of commodity . The location of real estate plays an important role in determining the price of dwelling. The determination of prices in a given area is closely related to the spatial distribution of properties. Some of the considered important factors are MRT LINE , Amenities like Hawker centers etc 

Data Sources:

          Transaction for private dwellings was fetched from Urban Development Authority of Singapore and also the MRT Line.

Address Geocoding:

           The location of a real estate is presented in the form of street name which should be converted to Lat Long points. We have used Python Geocode api for converting the street name to Latitude and Longitude .

Exploratory analysis:

        We have analysed total number of transaction taken place in each year for the dwelling with respect to a particular range of Unit Price.

styled-line(2)

Estimation of Distribution in Unit Real Estate Price:

 Spatial interpolation method Inverse distance weighting is used to generate models of selected dwellings. 

         The first issue appeared during data loading . Transactions of various dwelling located in same building is linked to the same address point and had a same position.Thus the average of values are taken to convert to a single value. Maps of the growth of dwelling values throughout the 2014-2017 characterized by a significant price increase.

Interpolation_2014_2015Interpolation_2016_2017

These maps shows that there is an increase in transactions around MRT lines and also significant price increase between 2016 and 2017 within unit price range of 2300 – 2600

Summary:

In particular, the use of a single color scale on the “interpolation map” allowed for the sharp increase in the unit prices of dwellings in late 2016 and 2017 to be captured, as also evidenced by graphs . Simultaneously, areas where the increase was found to be the highest and places that do not respond to trends in the market were indicated.

Team Name: HackEarth

Team Members: Navneeth Goswami , Vignesh Srinivasan , Praveen Kumar

Advertisements