Session-11/12

Apurva
30th Oct, 2017

setwd("~/Downloads/IIM Lucknow/TERM 5/DAM/Hotels")
library(readr)
hotel <- read.csv(paste("AllCitiesData.csv", sep=""))
summary(hotel)

    CityName      Population         CityRank     IsTouristDestination
 Jaipur : 765   Min.   :   8096   Min.   : 3.00   No :2168            
 Goa    : 624   1st Qu.: 595575   1st Qu.: 9.00   Yes:4299            
 Kochi  : 608   Median :1457723   Median :15.00                       
 Pune   : 600   Mean   :1515470   Mean   :16.03                       
 Udaipur: 456   3rd Qu.:2490891   3rd Qu.:23.00                       
 Agra   : 432   Max.   :4467797   Max.   :37.00                       
 (Other):2982                                                         
 IsWeekend  IsNewYearEve          Date     
 No :2424   No :5658     Dec 25 2016: 810  
 Yes:4043   Yes: 809     Dec 28 2016: 810  
                         Dec 18 2016: 809  
                         Dec 21 2016: 809  
                         Dec 24 2016: 809  
                         Dec 31 2016: 809  
                         (Other)    :1611  
                             HotelName       RoomRent       StarRating   
 Apnayt Villa                     :  16   Min.   :  299   Min.   :0.000  
 Fortune Hotel The South Park     :  16   1st Qu.: 2482   1st Qu.:3.000  
 Golden Tulip Chandigarh Panchkula:  16   Median : 4000   Median :3.000  
 Goldfinch Hotel                  :  16   Mean   : 6025   Mean   :3.424  
 Holiday Inn Chandigarh Panchkula :  16   3rd Qu.: 6499   3rd Qu.:4.000  
 Hotel City Park                  :  16   Max.   :75000   Max.   :5.000  
 (Other)                          :6371                                  
    Airport     
 Min.   : 0.20  
 1st Qu.: 6.20  
 Median :12.00  
 Mean   :15.17  
 3rd Qu.:22.50  
 Max.   :61.00  

                                                      HotelAddress 
 14-4507/9, Balmatta Road, Near Jyothi Circle, Hampankatta  :  16  
 57/1,Tamil Sangam Road, Simmakkal, Madurai, 625001, India  :  16  
 B-4/25, Shivala Ghat, Nepali Kothi, Varanasi, Uttar Pradesh:  16  
 Chowara Beach, Kovalam, Chowara, Kerala                    :  16  
 Circuit House Road,,Jodhpur,Rajasthan,India                :  16  
 Fatehabad Road, Agra, Uttar Pradesh, 282001 India          :  16  
 (Other)                                                    :6371  
  HotelPincode                                  HotelDescription
 Min.   :134109   3-star Jaipur hotel with restaurant   :  64   
 1st Qu.:282001   Abc                                   :  64   
 Median :342802   Jaipur hotel next to shopping         :  56   
 Mean   :388606   3 lmn                                 :  40   
 3rd Qu.:411057   3-star Jaisalmer hotel with restaurant:  40   
 Max.   :781022   3-star hotel in Pune                  :  40   
                  (Other)                               :6163   
 FreeWifi   FreeBreakfast HotelCapacity    HasSwimmingPool
 No : 542   No :2504      Min.   :  1.00   No :3867       
 Yes:5925   Yes:3963      1st Qu.: 15.00   Yes:2600       
                          Median : 30.00                  
                          Mean   : 49.55                  
                          3rd Qu.: 64.00                  
                          Max.   :414.00

1. How many total hotels are there in the sample?

hotel$id <- paste(hotel$HotelName, hotel$CityName)
length(unique(hotel$id))

[1] 821

2. How many total cities are there in the sample?

length(unique(hotel$CityName))

[1] 25

3. What is the average room rent in the sample?

mean(hotel$RoomRent)

[1] 6024.896

4. What is the median room rent in the sample?

median(hotel$RoomRent)

[1] 4000

Tourist vs. Non-Tourist Destinations

1a. List the Tourist destination cities in the sample

tourist <- subset(hotel, hotel$IsTouristDestination == "Yes")
length(unique(tourist$CityName))

[1] 13

1b. List the Non-tourist destination cities in the sample

nontourist <- subset(hotel, hotel$IsTouristDestination == "No")
length(unique(nontourist$CityName))

[1] 12

2. Create a table showing the count of hotels in Tourist and non-Tourist destinations

mt <- aggregate(id ~ IsTouristDestination, hotel, function(x) length(unique(x)))
colnames(mt) <- c("Tourist Destination","Count")
mt

  Tourist Destination Count
1                  No   292
2                 Yes   529

3. Create a table showing the average room rent in Tourist and non-Tourist destinations

mt1 <- aggregate(hotel$RoomRent ~ hotel$IsTouristDestination, FUN = mean)
colnames(mt1) <- c("Tourist Destination","Average Rent")
mt1

  Tourist Destination Average Rent
1                  No     4507.500
2                 Yes     6790.124

4. Plot the average rent in Tourist and non-Tourist destinations

library(gplots)
plotmeans(hotel$RoomRent ~ hotel$IsTouristDestination, mean.labels = TRUE, xlab = "Tourist Destination", ylab = "Average Room Rent")

plot of chunk unnamed-chunk-10

5. Compare the average room rent at Tourist destinations with non-Tourist destinations using a suitable statistical test

t.test(hotel$RoomRent ~ hotel$IsTouristDestination)


    Welch Two Sample t-test

data:  hotel$RoomRent by hotel$IsTouristDestination
t = -14.33, df = 6464.8, p-value < 2.2e-16
alternative hypothesis: true difference in means is not equal to 0
95 percent confidence interval:
 -2594.885 -1970.361
sample estimates:
 mean in group No mean in group Yes 
         4507.500          6790.124

5. Compare the average room rent at Tourist destinations with non-Tourist destinations using a suitable statistical test

t.test(hotel$RoomRent ~ hotel$IsTouristDestination)


    Welch Two Sample t-test

data:  hotel$RoomRent by hotel$IsTouristDestination
t = -14.33, df = 6464.8, p-value < 2.2e-16
alternative hypothesis: true difference in means is not equal to 0
95 percent confidence interval:
 -2594.885 -1970.361
sample estimates:
 mean in group No mean in group Yes 
         4507.500          6790.124

#As p-value < 0.05, we conclude that reject the Null hypothesis that average room rent is equal.

Effect of offering Swimming Pool on Hotel Rent at Tourist vs. Non-Tourist Destinations

1. Create a table counting the number of hotels offering and not-offering swimming pool at Tourist and non-Tourist destinations. Also include the total count of hotels

mytable2 <- with(hotel, tapply(id, list(IsTouristDestination,HasSwimmingPool), FUN = function(x) length(unique(x))))
colnames(mytable2) <- c("SP-No", "SP-Yes")
rownames(mytable2) <- c("TD-No", "TD-Yes")
mt2 <- subset(mytable2, select=c(2,1))
addmargins(mt2)

       SP-Yes SP-No Sum
TD-No      93   199 292
TD-Yes    231   300 531
Sum       324   499 823

2. Compare the proportion of hotels offering swimming pool at Tourist destinations with the proportion of hotels offering swimming pool at non-Tourist destinations, using an appropriate statistical test.

prop.table(mt2,1)

          SP-Yes     SP-No
TD-No  0.3184932 0.6815068
TD-Yes 0.4350282 0.5649718

3. Create a table summarizing the average room rent at hotels offering and not-offering swimming pool at Tourist and non-Tourist destinations.

a <- aggregate(hotel$RoomRent ~ hotel$IsTouristDestination + hotel$HasSwimmingPool, FUN = mean)
colnames(a)<- c("TouristDest","SwimmingPool","Average Rent")
a

  TouristDest SwimmingPool Average Rent
1          No           No     3328.812
2         Yes           No     3966.258
3          No          Yes     6691.175
4         Yes          Yes    10563.975

4. Visualize the differences using a suitable type of plot

a$lab <- c("NT-NoPool","T-NoPool","NT-Pool","T-Pool")
barplot(a$`Average Rent`, names.arg = a$lab, ylab = "AVerage Room Rent", col = rainbow(4))

plot of chunk unnamed-chunk-16

SESSION 12: Analysis of Impact of Destination & Hotel Capacity on Average Room Rent

TASK 4A: Is there statistical evidence to reject the following null hypothesis? H1: There is no influence of hotel capacity (i.e number of rooms in a hotel) on hotel rent

m0 <- lm(RoomRent ~ HotelCapacity, data = hotel)
summary(m0)


Call:
lm(formula = RoomRent ~ HotelCapacity, data = hotel)

Residuals:
   Min     1Q Median     3Q    Max 
 -8554  -3230  -2103    263  69933 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
(Intercept)   4824.668    125.213   38.53   <2e-16 ***
HotelCapacity   24.222      1.731   13.99   <2e-16 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 7334 on 6465 degrees of freedom
Multiple R-squared:  0.02938,   Adjusted R-squared:  0.02923 
F-statistic: 195.7 on 1 and 6465 DF,  p-value: < 2.2e-16

TASK 4B: Is there statistical evidence to reject the following null hypothesis? H3a: There is no influence of hotel capacity on room rent at non-Tourist destinations H3b: There is no influence of hotel capacity on room rent at Tourist destinations. How can we compare the differential impact of hotel capacity on room rent at Tourist versus non-Tourist destinations?

m2 <- lm(RoomRent ~ IsTouristDestination + HotelCapacity + IsTouristDestination:HotelCapacity, data = hotel)
summary(m2)


Call:
lm(formula = RoomRent ~ IsTouristDestination + HotelCapacity + 
    IsTouristDestination:HotelCapacity, data = hotel)

Residuals:
   Min     1Q Median     3Q    Max 
 -9176  -3249  -1623    428  69357 

Coefficients:
                                      Estimate Std. Error t value Pr(>|t|)
(Intercept)                           3439.566    215.215  15.982  < 2e-16
IsTouristDestinationYes               1885.360    263.152   7.165 8.67e-13
HotelCapacity                           18.903      2.638   7.165 8.67e-13
IsTouristDestinationYes:HotelCapacity   12.914      3.470   3.721    2e-04

(Intercept)                           ***
IsTouristDestinationYes               ***
HotelCapacity                         ***
IsTouristDestinationYes:HotelCapacity ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 7228 on 6463 degrees of freedom
Multiple R-squared:  0.05749,   Adjusted R-squared:  0.05706 
F-statistic: 131.4 on 3 and 6463 DF,  p-value: < 2.2e-16

Differential impact of hotel capacity on room rent at Tourist versus non-Tourist destinations

# As p-value of the interaction variable is less than 0.05, we conclude that the interaction effect between Tourist Destination and Hotel Capacity is statistically significant. This differential impact is captured by the magnitude of B3 = 12.9