Overview

As posted by Donghwan Kim, the analysis is based on the data on students living on and off campus source.

Read the CSV dataset

Step 1 is to load the csv from the github library and to load the required R libraries.

knitr::opts_chunk$set(eval = TRUE, results = FALSE)
library(tidyverse) 
library(RCurl)
library(stringr)
library(knitr)
library(openxlsx)

campus_raw_url <- getURL("https://raw.githubusercontent.com/BharaniNittala/DATA-607/master/Proj2_Case1.csv") 
campus_raw <- read.csv(text = campus_raw_url,na.strings=c("","NA"))

knitr:: kable(campus_raw)
State.of.residence X X.1 Do.you.live.on.campus. X.2 Total
NA NA NA Off-campus On-campus NA
In state Class Rank Underclassman 58 110 168
NA NA Upperclassman 108 7 115
NA Total NA 166 117 283
Out of state Class Rank Underclassman 13 30 43
NA NA Upperclassman 39 2 115
NA Total NA 52 32 158
Total Class Rank Underclassman 71 140 211
NA NA Upperclassman 147 9 156
NA Total NA 218 149 367

Tidy the data with tidyr

names <- c("Residence","Class_Rank","Class","Off_campus","On_campus","Total")
colnames(campus_raw) <- names

#Selecting only the columns that are required and applying fill function to impute missing values
campus_raw_v2 <- campus_raw[c(-1,-4,-7,-10),c(-2,-6)]
campus_raw_v2 <- campus_raw_v2 %>% fill(Residence)


campus_tall <- pivot_longer(campus_raw_v2,names[4:5],names_to = "Campus", values_to = "Students",values_drop_na = TRUE)

Analysis

Write code to quantify change in student’s residency as class rank goes up (by state and by on- and off-campus)

campus_tall$Students <- as.numeric(as.character(campus_tall$Students))
campus_tall_v2 <- campus_tall[,-3] %>% group_by(Class,Residence) %>% summarise(Grouped_students = sum(Students))
campus_tall_v2_stat <- pivot_wider(campus_tall_v2, names_from = "Residence", values_from = "Grouped_students")

campus_tall_v2_stat <- mutate(campus_tall_v2_stat, "OOS_pct" = round(( `Out of state` / Total ) * 100, 2), "IS_pct" = round(( `In state` / Total ) * 100, 2) ) 


dodger = position_dodge(width = 0.9)
ggplot(campus_tall_v2_stat, aes(y=OOS_pct, x=Class,  fill = Class)) + 
    geom_bar( stat="identity",position = dodger, color="black")+
    geom_text(aes(label=OOS_pct),color = "blue",position = dodger,vjust=-0.5)+
  scale_fill_manual(values = alpha(c("#000000", "#FF5733"))) +
  ylab("Percentage of Out of State")

dodger = position_dodge(width = 0.9)
ggplot(campus_tall_v2_stat, aes(y=IS_pct, x=Class,  fill = Class)) + 
    geom_bar( stat="identity",position = dodger, color="black")+
    geom_text(aes(label=IS_pct),color = "blue",position = dodger,vjust=-0.5)+
  scale_fill_manual(values = alpha(c("#000000", "#FF5733"))) +
  ylab("Percentage of In State")

We see that as the class progresses (underclass to upperclass), more studets stay out of state

Find Student’s choice pattern of on- and off-campus by state and class rank.

campus_tall_v3 <- campus_tall %>% group_by(Class,Residence,Campus) %>% summarise(Grouped_students = sum(Students))
campus_tall_v3_stat <- pivot_wider(campus_tall_v3, names_from = "Residence", values_from = "Grouped_students")
campus_tall_v3_stat <- mutate(campus_tall_v3_stat, "OOS_pct" = round(( `Out of state` / Total ) * 100, 2), "IS_pct" = round(( `In state` / Total ) * 100, 2) ) 

dodger = position_dodge(width = 0.9)
ggplot(campus_tall_v3_stat, aes(y=OOS_pct, x=Class, color=Campus, fill=Campus)) + 
    geom_bar( stat="identity",position = dodger, color="black")+
    geom_text(aes(label=OOS_pct),color = "blue",position = dodger,vjust=-0.5)+
  scale_fill_manual(values = alpha(c("#000000", "#FF5733"))) +
  ylab("Percentage of Out of State")

dodger = position_dodge(width = 0.9)
ggplot(campus_tall_v3_stat, aes(y=IS_pct,x =Class, color=Campus, fill=Campus)) + 
    geom_bar( stat="identity",position = dodger, color="black")+
    geom_text(aes(label=IS_pct),color = "blue",position = dodger,vjust=-0.5)+
  scale_fill_manual(values = alpha(c("#000000", "#FF5733"))) +
  ylab("Percentage of In State")

We see that as the class progresses (underclass to upperclass), more studets stay out of state from previous graph. Now, when we look deeper into the ‘underclass ’Out of State’, we see that underclassman prefer on-campus more than off-campus and reverse trend for upperclassman where they prefer off-campus. When we look into ‘In State’ residence, Underclassman slightly prefer off-campus more than on-campus compared to upperclassman where their preference is to stay on-campus when ‘In state’.

Conclusion

We see interesting choice of on-campus and off-campus when in-state and out of statement among the underclassman and upperclassman. I would have combined this dataset with tuition fees and hostel fees to understand the behavior furthermore.