ISSS608 2016-17 T1 Assign1 Abhinav Ghildiyal

Abstract

Singapore has a land mass of 719.1 square km with over 6 million people, and this is the reason why Singapore is one of the worlds most expensive housing market. Singapore residential property falls into 2 categories : Public housing (HDB flats) and Private Housing (Condominiums, Landed property and Executive Condominiums). In this write up we will discuss over the resale pricing of the Public Housing and we will try to analyse the price variation over the years based on different underlying factors such as flat type, month, flat model etc.

Problem and Motivation

This write-up will be a help for buyers/sellers to observe, analyse and decide the following concerns:

1. When to buy/sell the property?
2. Which flat type to buy/sell?
3. Which flat model to buy/sell?

Tools Used

Tableau
JMP Pro
MS Excel

Approach

I have taken the data from the Singapore Government Data Portal (https://data.gov.sg/) - the government`s one stop portal to its publicly-available data sets from 70 public agencies - and used different visualsing tools to make the infograhics and deduce the concerns mentioned above. The first step here is to determine whether there is any significant difference between the means of the different flat types, because here I am using the sample of flat type and the number of resale application to analyse the variation.

Notice that the flat_types do not spread uniformly across the horizontal axis. The dot plot of flat_types = 4 - room indicates that there are many more no_of_resale_applications than the other 5 flat_types, whereas the dot plot of flat_types = 1 – room indicates that it has the lowest no_of_resale_applications.

Mean Comparison using ALL Pair,Tukey HSD test

Now if we do the mean comparison using the ALL Pair, Tukey HSD to test the difference in the mean of different flat types, by examining how the comparison circles interacts, it is quite prominent that there is the significant difference in the means of all the flat_types.

To showcase the share of resale public housing supply in 2015, I have used number-of-resale-applications-registered-by-flat-types data set from Data.gov.sg website. This data set has the count of no_of_resale_application by flat_types and quarter. Using Tableau, I created the distribution of no_of_resale_applications by flat_type and quarter, below is the distribution view.

Distribution of Number of resale applications by flat types and quarter

But we need the share of resale public housing supply in 2015, so I have used the filter operation on Quarter and then selected Q1, Q2, Q3 and Q4 of year 2015.

The below Pareto chart shows that flat_type = 4 Rooms covers 39.66% (i.e. 7656 resale application) of the total no_of_resale_application (i.e. 19306 resale applications), in the similar fashion it will tell the details of all the flat_types. From this visualization we can make the inference that the max number of resale application has come for flat type = 4 room that means:
Either demand is more for the flat type = 4 Room or there is a rise in the economy and people are not able to afford the 4 Room flats but to make a final conclusion we need several other factors such as inflation rate, buyers market etc.

RESULT

Distribution of Number of resale applications by flat types for the year 2015

                                              The table above shows the percentage and the count of the number of the resale applications based on flat type for the year 2015.

Distribution of Number of resale applications by flat types for Quarter 1 of 2015

Now if we want to visualise the distribution of No of resale applications by flat types for particular quarters of year 2015, we can select the particular quarter and hover the mouse to see the details. Here we can see that there are 925 no_of_resale_applications for flat_type = room 5 for quarter 1 of 2015. In the similar fashion we can see the no_of_resale_applications by hovering the mouse and selecting different points and area in the graph.

Quarter that has maximum number of resale applications

The table has been extracted from the Tableau distribution chart (Distribution of No of resale applications by flat types for year 2015). Here we can observe that flat_type = 5 room, 4 room, 3 room and 1 room has the max number of resale applications in the 2nd quarter of 2015, whereas flat_type = Executive and 2 room has their max number of resale application in the 4th quarter of the year 2015. From the distribution chart we can figure out that flat_type = 4 rooms have the maximum number of resale applications i.e. 7656 resale applications (39.66%) of the total 19306 resale applications.

Resale Price of flats by flat type

This distribution graph shows the Resale price of the flats based on the type of flats for the year 2015. From this graph we can infer that the sum of resale price of Flat_type = 4 Rooms is the highest.

Flat model with highest price

From this distribution we can infer that in a year, Flat Model = Model A for 10 months has the highest sum of resale price, where as in other 2 months Flat Model = Improved has the highest resale price.

Resale price by flat type and flat model

The box and whisker plot can help us see data through month and flat types. The upper and lower whisker and the median can be displayed in the graph. Here we can infer that the sum of resale price is highest in the month of October for 4 Room Model A flats.

Comparing the resale price by flat type for 1st quarter of the year 2015 with the resale price by flat type for 1st quarter of the year 2016

From this comparison we can infer that the sum of price for the 4 Room flat in 1st quarter of 2016 has increased if compared with the resale price of 1st quarter of year 2015. It means that there is a uptrend and more people are interested in buying the 4 room flats. Infact, there is an increase in the resale price for almost all kind of flat type. It shows that overall buying capacity of people from 2015 to 2016 has increased to a certain level.

INFOGRAPHICS