ISSS608 2016-17 T1 Assign1 Kanokkorn

Problem and Motivation

Purchasing a room is a big decision which involves large amount of money and long-term commitment. Therefore, purchaser needs to consider information in all aspects before making purchase decision. This study aimed to provide the overview of public housing flat resale in 2015 and first-half of 2016.

Approaches

1. Data Acquiration

The first step to meet the goal was to acquire data. The data was found at Government database (https://data.gov.sg/dataset/resale-flat-prices). Concerned with the price in 2015 and 2016, we used the file named “Resale Flat Prices (Based on Registration Date), From March 2012 Onwards” which covered from March 2012 to June 2016 [as stated in the website].

2. Data Transformation

The data given was quite clean already when importing the text file to Tableau. There was only one transformation of ‘Month’/’Lease commence date’ to be data-time format instead of numeric / text. The transformation is to create calculated field by using ‘DATE’ function and text join (In Tableau, joining text uses ‘+’ instead of ‘&’ in excel). For example, we can transform month field using DATE(""+left([Month],4)+"-"+RIGHT([Month],2)+"-01").

3. Data Analysis

Only descriptive analysis techniques were used in this project.

Tools Utilised

Tableau

Only tableau is used as data analysis tool from data cleaning to output.

Result

Flat Type

In 2015, the most common flat type resold is 4-room. Followed by 3-room and 5-room. The trend was the same in 2015 and in the first-half of 2016. However, looking at Singapore average household size in 2015 was only 3.37 (http://www.singstat.gov.sg/statistics/latest-data). Does this mean that Singaporean are sharing flat with others?

Floor

When it comes to which floor the flat is located, with more supply from 3-room, 4-room, and 5-room flat, these group provides buyer choices of both high floor and lower floor units. Also, there are plenty of lower floor units from this group of units offered. The trend is similar in the first half of 2016 and 2015.

Flat Town

For the town of units in 2015 based on transaction value, 80% of the records came from top 32 towns (out of total 52 towns) the shares of township did not seem to follow ‘80/20 Pareto rule’.

The distribution of the town showed the increase of flat sold shares in Sengkang, Choa Chu Kang, and Punggol.

Resale Price

The distribution of resale price was positively-skewed for both 2015 and 2016. Looking at each of the town, some had wide spread such as Central Area and some had narrower spread such as Sembawang. Central Area was where the highest median of resale price located and had largest spread.

Central Area : Price Distribution from Streets

Drilling down the distribution of resale price from each street in central area, large amount of transaction from ‘Cantonment Road’ (which was resold at significantly higher than other areas) was the reason why the median of Central Area was high as well as the mean.