The Melbourne Housing And Its Pricing.
Sharmistha Sutradhar
January 31, 2018
1. Introduction
Melbourne is the state capital and most populous city of the Australian state of Victoria, and the second-most populous city in Australia. Melbourne rates highly in education, entertainment, health care, research and development, tourism and sport making it the world’s most liveable city. Melbourne has minimal public housing and high demand for rental housing, which is becoming unaffordable for some.Public housing is usually provided by the Housing Commission of Victoria, and operates within the framework of the Commonwealth-State Housing Agreement, by which federal and state governments provide housing funding. But when it comes to properties, and its buying and selling, it becomes a very crutial issue. This is because a property’s price reflects an assessment of the value that the property has based on the locality and various other factors, and also their(buyers) willingness-to-pay for the property. This analysis is focussed on the “pricing of the property”, based on the quality of accomodation it provides.
2. Overview of the study
Our research study concerns property/housing prices in Melbourne which is located in Australia. It is the beating heart of the city and a magnet for everyone who surrounds it. Melbourne is experiencing high population growth, generating high demand for housing. This housing boom has increased house prices and rents, as well as the availability of all types of housing. Subdivision regularly occurs in the outer areas of Melbourne, with numerous developers offering house and land packages. However, after 10 years of planning policies to encourage medium-density and high-density development in existing areas with greater access to public transport and other services, Melbourne’s middle and outer-ring suburbs have seen significant brownfields redevelopment. Based on strategies, the prices of the housing are decided based on the numbers of rooms and on other accomodation factors. We empirically study how the factors like distance from the housing, number of rooms, landsize and car-parking areas influence the housing prices.
3. An empirical analysis of The Melbourne Housing prices at Melbourne, Australia.
3.1 Overview
The main motive behind this Study was to investigate the pricing strategy employed by housingslocated in the city of Melbourne. This study analyzed property prices at Melbourne, Australia. Our goal was to compare prices of housings on the basis of the number of rooms, with the parking area and landsize. The rationale behind this is summarized next.
Hypothesis H1: There is no significant difference in the prices of the housing which are having more landsize than those which occupy less area.
3.2 Data
To conduct this Study, the data was collected from the website (https://www.kaggle.com/datasets). The Housing societies in the city of Melbourne has a lot of various. Beacuse this data was collected only from a few areas, so the sellers are common for most of it. The Housing Society has a lot of variations, not excatly variations but can be called types. The properties differ on the basis of the “number of rooms” it has and “how many bedrooms and bathrooms” it has and “the landsize of the property”.
Locality: Locality, that is, the place where the societies are situated are highly important because that will decide the kind of buyers of those housing. The finanial status of the customers will decide with housing they’d prefer according to their suitable locality. Locality also refers to the connectivity of that place to the rest of the city with ease.
Family Size: The family size also depends on how big or small a property you want to buy. again, if the family is huge, it might have kids and working ppeople who will require good and efficient means to commute from one place to another, connectivity with schools and shops around.
3.3 Model
In order to test Hypothesis 1a, we proposed the following model:
Price ~ Rooms+Landsize+Bedroom+Bathroom+Car
regg<- lm(MHousing.df$Price~MHousing.df$Rooms+MHousing.df$Landsize+MHousing.df$Bedroom2+MHousing.df$Bathroom+MHousing.df$Car)
summary(regg)##
## Call:
## lm(formula = MHousing.df$Price ~ MHousing.df$Rooms + MHousing.df$Landsize +
## MHousing.df$Bedroom2 + MHousing.df$Bathroom + MHousing.df$Car)
##
## Residuals:
## Min 1Q Median 3Q Max
## -3074991 -323854 -106127 229091 9557236
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 68734.805 15027.355 4.574 4.82e-06 ***
## MHousing.df$Rooms 220485.592 14835.579 14.862 < 2e-16 ***
## MHousing.df$Landsize 2.220 1.121 1.980 0.0478 *
## MHousing.df$Bedroom2 -14473.980 14602.386 -0.991 0.3216
## MHousing.df$Bathroom 243346.872 8000.809 30.415 < 2e-16 ***
## MHousing.df$Car 8416.958 4932.561 1.706 0.0880 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 560000 on 15760 degrees of freedom
## (15165 observations deleted due to missingness)
## Multiple R-squared: 0.2621, Adjusted R-squared: 0.2619
## F-statistic: 1120 on 5 and 15760 DF, p-value: < 2.2e-16t.test(MHousing.df$Price, MHousing.df$Landsize)##
## Welch Two Sample t-test
##
## data: MHousing.df$Price and MHousing.df$Landsize
## t = 253.9, df = 24198, p-value < 2.2e-16
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## 1043882 1060125
## sample estimates:
## mean of x mean of y
## 1052590.4437 587.0988We here see the effect of the above mentioned factors on the price of a property with the simplest regression model.
3.4 Results
From the regression analysis, we see that the price of the Melbourne housing is greatly affected by the number of rooms in total, especially The nuber of bathrooms as it is a great necessity. The landsize undoubtly matters for the buys because it gives him the value of asset he has bought.
Also through this t-test we reject our null hypothesis becuase the p-value is less than 0.05, which tells us that there is a significant diiference in the prices of the properties which have different landsizes. If we look at the regression model, we see that the price will increase by 1$, if the landsize changes by 2.2 unit square.
4. Conclusion
This paper was motivated by the need for research that could improve our understanding building patterns influences the pricing strategies in a housing society. The attrative research was that to show, prices actually influneces on many factors. It is just not the choice of a seller to set a price. There are strategies an importnat though-process which undergoes in deciding the prices of the housing.
5. References
https://www.kaggle.com/datasets https://en.wikipedia.org/wiki/Melbourne#Housing https://www.udemy.com/data-science-and-analytics-using-r/learn/v4/overview
6. Summary Statistics in the Melbourne Housing Prices
library("psych", lib.loc="~/R/win-library/3.4")
describe(MHousing.df[,c(3,5,9,11:15)])## vars n mean sd median trimmed mad
## Rooms 1 30931 3.02 0.97 3 2.99 1.48
## Price 2 24197 1052590.44 644499.49 875000 953681.82 437367.00
## Distance* 3 30931 117.62 77.58 101 117.00 109.71
## Bedroom2 4 23868 3.05 0.99 3 3.02 1.48
## Bathroom 5 23862 1.61 0.72 1 1.50 1.48
## Car 6 23497 1.70 1.00 2 1.61 1.48
## Landsize 7 20405 587.10 3540.11 500 451.57 329.14
## BuildingArea 8 12397 158.34 418.09 133 141.88 59.30
## min max range skew kurtosis se
## Rooms 1 16 15 0.54 2.75 0.01
## Price 85000 11200000 11115000 2.59 13.33 4143.26
## Distance* 1 232 231 0.09 -1.62 0.44
## Bedroom2 0 30 30 1.52 28.69 0.01
## Bathroom 0 12 12 1.42 5.30 0.00
## Car 0 26 26 2.14 22.71 0.01
## Landsize 0 433014 433014 95.26 11096.44 24.78
## BuildingArea 0 44515 44515 96.90 10223.76 3.767.One-way table for the variable: Type.
table23<-table(MHousing.df$Type)
table23##
## h t u
## 21160 3178 6593prop.table(table23)*100##
## h t u
## 68.41033 10.27448 21.315198. Two-way table for the variables: Type and Method.
table22<-xtabs(~MHousing.df$Type+MHousing.df$Method)
table22## MHousing.df$Method
## MHousing.df$Type PI PN S SA SN SP SS VB W
## h 2733 155 12555 138 944 2896 23 1630 86
## t 533 40 1693 24 108 446 4 319 11
## u 949 69 3439 40 171 1161 4 707 53prop.table(table22)*100## MHousing.df$Method
## MHousing.df$Type PI PN S SA
## h 8.83579580 0.50111539 40.59034625 0.44615434
## t 1.72319033 0.12932010 5.47347321 0.07759206
## u 3.06811936 0.22307717 11.11829556 0.12932010
## MHousing.df$Method
## MHousing.df$Type SN SP SS VB
## h 3.05195435 9.36277521 0.07435906 5.26979406
## t 0.34916427 1.44191911 0.01293201 1.03132779
## u 0.55284343 3.75351589 0.01293201 2.28573276
## MHousing.df$Method
## MHousing.df$Type W
## h 0.27803821
## t 0.03556303
## u 0.171349139. Boxplot of the concerned variables.
boxplot for price.
boxplot(MHousing.df$Price, xlab="Price of the property",ylab="Melbourne Housing", horizontal=TRUE)
boxplot for the number of rooms.
boxplot(MHousing.df$Rooms, xlab="Number of rooms",ylab="Melbourne Housing", horizontal=TRUE)
boxplot for the no. of bedrooms.
boxplot(MHousing.df$Bedroom2, xlab="Number of bedrooms",ylab="Melbourne Housing", horizontal=TRUE)
boxplot for the number of bathrooms.
boxplot(MHousing.df$Bathroom, xlab="Number of bathrooms",ylab="Melbourne Housing", horizontal=TRUE)
boxplot for the number of cars.
boxplot(MHousing.df$Car, xlab="Number of cars",ylab="Melbourne Housing", horizontal=TRUE)
boxplot for the size of the land.
boxplot(MHousing.df$Landsize, xlab="Landsize of the property",ylab="Melbourne Housing", horizontal=TRUE)
10. Corrgram of all the variables.
library("corrgram", lib.loc="~/R/win-library/3.4")
corrgram(MHousing.df, order=TRUE, lower.panel=panel.shade,
upper.panel=panel.pie, text.panel=panel.txt,
main="Corrgram of Variables")
11. Correlation between variables.
cor.test(MHousing.df$Price, MHousing.df$Rooms)##
## Pearson's product-moment correlation
##
## data: MHousing.df$Price and MHousing.df$Rooms
## t = 82.907, df = 24195, p-value < 2.2e-16
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
## 0.4604913 0.4801167
## sample estimates:
## cor
## 0.4703622cor.test(MHousing.df$Price, MHousing.df$Landsize)##
## Pearson's product-moment correlation
##
## data: MHousing.df$Price and MHousing.df$Landsize
## t = 4.2764, df = 15944, p-value = 1.91e-05
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
## 0.01833626 0.04934331
## sample estimates:
## cor
## 0.0338479312. Scatterplot of price and number of rooms.
library("car", lib.loc="~/R/win-library/3.4")##
## Attaching package: 'car'## The following object is masked from 'package:psych':
##
## logitscatterplot(MHousing.df$Price,MHousing.df$Rooms,xlab = "price", ylab = "number of rooms")
13. Scatterplot of price and landsize.
library("car", lib.loc="~/R/win-library/3.4")
scatterplot(MHousing.df$Price,MHousing.df$Landsize,xlab = "price", ylab = "landsize")