Sampath Rathnayake, Melbourne, Victoria
Last updated: 19 November, 2017
There are many hotels, clubs and other recreational facilities in Victoria that have electronic gaming machines installed and we can see there are many people expending their money gambling in these machines. It is important to explore the patterns of peoples’ expenditure on these machines for social and economical aspects.
This study tries to explore the patterns of expenditure based on the venue with associated patameters venue type, local goverment area and region in Victoria.
In this study we mainly focus on,
Dataset used in this analysis was sourced from Victoria’s open data directory.
https://www.data.vic.gov.au/data/dataset/current-gaming-expenditure-by-venue
The agency program of this dataset was Victorian Commission for Gambling and Liquor Regulation under the Department of Justice and Regulation.
This data set provides information relating to the total expenditure and gaming venue information including venue classification and the allocation of electronic gaming machines (EGMs) throughout Victoria.
Regions: Gaming venues are classified one of two regions, Country or Metro.
LGAs : Totally 70 Local Government Authorities covering all in Victoria.
Venue Types: Gaming venues are classified one of two types, Hotel or club.
Expenditure: Amount of money lost by gaming patrons. Also referred to as ‘player loss’. Expenditure is in AUD millions.
| Region | Min | Q1 | Median | Q3 | Max | Mean | SD | n | Missing |
|---|---|---|---|---|---|---|---|---|---|
| Country | 0.245 | 1.3895 | 2.339 | 4.1870 | 11.253 | 3.008775 | 2.200936 | 187 | 0 |
| Metro | 0.102 | 3.2790 | 5.399 | 9.1175 | 20.963 | 6.498032 | 4.248518 | 315 | 0 |
| Ven_type | Min | Q1 | Median | Q3 | Max | Mean | SD | n | Missing |
|---|---|---|---|---|---|---|---|---|---|
| Club | 0.102 | 1.7485 | 3.186 | 5.1220 | 16.059 | 3.783151 | 2.777681 | 239 | 0 |
| Hotel | 0.273 | 2.9270 | 5.472 | 9.4555 | 20.963 | 6.484213 | 4.475324 | 263 | 0 |
\[H_0: \mu_1 = \mu_2 \] \[H_A: \mu_1 \ne \mu_2\] - In order to use the t-test sampling distribution of data should be normal. Here we are going to test the normality assumption of two samples (Country and Metro) 
- The normality assumption is Violated for both samples. However, our samples are large enough to ignore this assumption under the Central Limit Theorem - Testing the assumption of Homogeneity of Variance of sampling populations
##
## Welch Two Sample t-test
##
## data: Expenditure by Region
## t = -12.096, df = 492.24, p-value < 2.2e-16
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -4.056009 -2.922504
## sample estimates:
## mean in group Country mean in group Metro
## 3.008775 6.498032The small p-value provides strong evidence to reject the null hypothesis in favour of alternative hypothesis suggesting that the average expenditure for two areas is different and as we saw in summary statistics the average expenditure or player loss for Metro is greater than Country.
##
## Welch Two Sample t-test
##
## data: Expenditure by Ven_type
## t = -8.2025, df = 443.48, p-value = 2.569e-15
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -3.348239 -2.053886
## sample estimates:
## mean in group Club mean in group Hotel
## 3.783151 6.484213The small p-value provides strong evidence to reject the null hypothesis in favour of alternative hypothesis suggesting that the average expenditure for two venues is different and as we saw in summary statistics the average expenditure or player loss for Hotels is greater than Clubs.
-In order to test whether there’s any association in the population between LGA and Venue Type we use below hypothese, \[H_A: \] There is no association in the population between LGA and Venue Type Vs. \[H_A: \] There is an association in the population between LGA and Venue Type
##
## Club Hotel
## Borough of Queenscliffe 1 0
## City of Ballarat 7 8
## City of Banyule 5 5
## City of Bayside 1 4
## City of Boroondara 0 4
## City of Brimbank 6 9
## City of Casey 4 9
## City of Darebin 3 9
## City of Frankston 3 6
## City of Glen Eira 6 5
## City of Greater Bendigo 5 6
## City of Greater Dandenong 9 6
## City of Greater Geelong 12 14
## City of Greater Shepparton 4 4
## City of Hobsons Bay 8 2
## City of Hume 5 9
## City of Kingston 8 8
## City of Knox 4 7
## City of Latrobe 9 4
## City of Manningham 3 4
## City of Maribyrnong 4 5
## City of Maroondah 6 4
## City of Melbourne 4 7
## City of Monash 5 10
## City of Moonee Valley 6 6
## City of Moreland 6 6
## City of Port Phillip 2 8
## City of Stonnington 3 4
## City of Warrnambool 4 4
## City of Whitehorse 2 4
## City of Whittlesea 4 6
## City of Wyndham 6 7
## City of Yarra 1 7
## Rural City of Ararat 2 0
## Rural City of Benalla 2 1
## Rural City of Horsham 3 0
## Rural City of Mildura 6 3
## Rural City of Swan Hill 3 1
## Rural City of Wangaratta 2 2
## Rural City of Wodonga 0 3
## Shire of Alpine 1 1
## Shire of Bass Coast 4 1
## Shire of Baw Baw 2 2
## Shire of Campaspe 2 2
## Shire of Cardinia 2 3
## Shire of Central Goldfields 2 0
## Shire of Colac-Otway 3 2
## Shire of Corangamite 1 1
## Shire of East Gippsland 8 2
## Shire of Gannawarra 1 0
## Shire of Glenelg 2 2
## Shire of Hepburn 1 1
## Shire of Macedon Ranges 2 1
## Shire of Mansfield 1 0
## Shire of Melton 3 4
## Shire of Mitchell 2 3
## Shire of Moira 1 1
## Shire of Moorabool 2 1
## Shire of Mornington Peninsula 6 11
## Shire of Mount Alexander 0 1
## Shire of Murrindindi 0 1
## Shire of Nillumbik 0 2
## Shire of Northern Grampians 2 0
## Shire of South Gippsland 2 2
## Shire of Southern Grampians 1 1
## Shire of Strathbogie 1 0
## Shire of Surf Coast 1 2
## Shire of Towong 1 0
## Shire of Wellington 6 1
## Shire of Yarra Ranges 5 4##
## Pearson's Chi-squared test
##
## data: table(Assignment4_data$LGAName, Assignment4_data$Ven_type)
## X-squared = 67.972, df = 69, p-value = 0.5124The large p-value (0.5124> 0.05) does not provide evidence to reject null hypothesis in favour of alternative hypothesis and therefore there is no statistically significant association between LGA and Venue Type. The visualised association (hotels> Clubs) is only due to natural variability of the sample.