Top World Universities 2024
Hannah Hodges
2024-02-20
Introduction
The chosen data set was chosen by someone who is a current college student and I am interested to see how these top universities across the world are scored. It would also be interesting to see if someone could use this data set when thinking about which school they may want to attend. In this report, we will be focusing on the overall score, employment outcomes, international student ratio, international faculty ratio, international research network, and academic reputation. A majority of the visuals will look at how these variables relate to the overall score. From this, I hope to find which variables contribute most to the overall score in the data set.
Dataset
Description
The data set used is the Top 100 Universities across the world based on several indicators. This data set was chosen to see which universities rank across the world. The columns include Rank, University Name, Overall Score, Academic Reputation, Employer Reputation, Faculty to Student Ratio, Citations per faculty, International Faculty Ratio, International Student Ratio, International Research Network, Employment Outcomes, and Sustainability. All columns hold numerical data except the University name. The score is scored out of 100. Rank is based on which university out of the 100, scored the highest out of all the indicators. Academic Reputation score is given based on the distinction of academics. Employer Reputation is a score based on the how how employees see the university as a place of work. Faculty to Student Ratio is the percentage of faculty to students. Citations per faculty is based on the number of citations acknowledged in papers over a five-year period by faculty members at said institution. International Faculty Ratio is based on how many International faculty members there are. International Student Ratio is the same as the International Faculty Ratio except with students. International Research Network has to do with the community of researchers that the university has. Employment outcome score is based on how students place jobs after graduation. Lastly, sustainability is due to the university’s devotion to environmental sciences.
Stats
Before we go into the report, let’s look at some descriptive stats:
#setting our dataset
top_100_world_university_2024 <- read.csv("//apporto.com/dfs/LOYOLA/Users/hmhodges_loyola/Desktop/IS 460 Data/top 100 world university 2024.csv")
#This provides the descriptive stats of our dataset
summary(top_100_world_university_2024)## X rank university overall_score
## Min. : 0 Length:105 Length:105 Min. : 59.40
## 1st Qu.: 26 Class :character Class :character 1st Qu.: 64.70
## Median : 52 Mode :character Mode :character Median : 73.40
## Mean : 52 Mean : 75.02
## 3rd Qu.: 78 3rd Qu.: 84.50
## Max. :104 Max. :100.00
## academic_reputation employer_reputation faculty_student_ratio
## Min. : 43.40 Min. : 27.30 Min. : 4.40
## 1st Qu.: 71.30 1st Qu.: 59.90 1st Qu.: 34.90
## Median : 85.00 Median : 84.70 Median : 61.30
## Mean : 82.72 Mean : 77.49 Mean : 59.84
## 3rd Qu.: 97.60 3rd Qu.: 96.30 3rd Qu.: 90.40
## Max. :100.00 Max. :100.00 Max. :100.00
## citations_per_faculty international_faculty_ratio international_students_ratio
## Min. : 1.70 Min. : 6.00 Min. : 2.30
## 1st Qu.: 51.40 1st Qu.: 39.10 1st Qu.: 45.70
## Median : 72.40 Median : 88.90 Median : 78.80
## Mean : 70.59 Mean : 70.16 Mean : 67.96
## 3rd Qu.: 93.50 3rd Qu.: 99.10 3rd Qu.: 96.80
## Max. :100.00 Max. :100.00 Max. :100.00
## international_research_network employment_outcomes sustainability
## Min. : 1.30 Min. : 13.40 Min. : 13.40
## 1st Qu.: 58.30 1st Qu.: 39.60 1st Qu.: 71.80
## Median : 85.20 Median : 66.00 Median : 83.70
## Mean : 74.53 Mean : 63.66 Mean : 78.72
## 3rd Qu.: 95.50 3rd Qu.: 92.50 3rd Qu.: 95.10
## Max. :100.00 Max. :100.00 Max. :100.00Findings
Bar Chart
#This will read our dataset
top_100_world_university_2024 <- read.csv("//apporto.com/dfs/LOYOLA/Users/hmhodges_loyola/Desktop/IS 460 Data/top 100 world university 2024.csv")
library(ggplot2)## Warning: package 'ggplot2' was built under R version 4.3.2#Bar Chart of overall score of the universities
#Using ggplot to create the bar chart
#We are using the overall_score as the x variable and count is our y
ggplot(top_100_world_university_2024, aes(x = overall_score, fill = after_stat(count)))+
geom_bar()+
#This sets the colors
scale_fill_gradient(low = "blue", high = "red")+
labs(title = "Overall Scores of Universities", x= "Overall Score", y= "Number for each Score")
As you will find, the scores are spread across the range of scores. A majority of the university scores are alone in their score, however, there are a few that have 2 or 3 with the same score. There is a single university that scored 100.
Corrplot
Moving on to the next visual. We will now be looking at a visual called a correlogram. This will show the correlation between the variables. In a correlogram, the intensity of the color and size of the circles represent the strength of correlation between pairs of variables. The darker and larger circles are, the stronger correlations will be. The lighter and smaller circles indicate weaker correlations.
library(corrplot)## Warning: package 'corrplot' was built under R version 4.3.2## corrplot 0.92 loaded#Calling our data
data <- top_100_world_university_2024[,-1]
#We are making sure all of our data is numeric
numeric_data <- data[, sapply(data, is.numeric)]
correlation_matrix <- cor(numeric_data)
corrplot(correlation_matrix, method = "circle", tl.cex = 0.7)
In this visual, we want to focus on the relationships and patterns within the data set. Therefore, with the provided correlogram, we can see all of the variables. From this we can see, which variables have strong correlations and which have weak correlations. One of our main focuses in this data set is to see which variables contribute most to the overall score.This visual is interesting because you can see the weak correlation and the very strong ones.
Correlation
Next we are going to look more closely at the individual variables. We are going to look at academic reputation and employment outcomes. To look at these variables we are going to look at the correlation between these two variables.Before looking at the results, one may believe that academic reputation would have a strong positive correlation to employment outcomes. For example, the stereotype of going to an IVY League, i.e. Harvard, Yale, Brown, etc, would place one at a higher advantage because of the institution’s academic reputation. Let see what it means for this data set.
#Here we are picking our x and y variables
ggplot(top_100_world_university_2024, aes(x = academic_reputation, y= employment_outcomes))+
geom_point()+
#Setting the color
geom_smooth(method = "lm", se= FALSE, color = "red")+
labs(title = "Scatter Plot of Academic Reputation vs. Employment Outcomes", x= "Academic Reputation", y= "Employment Outcomes")+
annotate("text", x = max(top_100_world_university_2024$academic_reputation), y = min(top_100_world_university_2024$employment_outcomes),
#Labeling our correlation outcome
label = paste("Correlation:",
#This provides the correlation data
round(cor(top_100_world_university_2024$academic_reputation, top_100_world_university_2024$employment_outcomes), 2)),
vjust = 1, hjust =1)## `geom_smooth()` using formula = 'y ~ x'
From this visual, we can see that there is a 53% correlation between academic reputation and employment outcomes in the data set. This visual is slightly surprising because before the correlation was run, one would think that there would be a high correlation between the two relating to the overall score. However, for the data in this data set, there are could be other variables that have a higher correlation.
Density Plot
Now lets look at the next visual. Below you will find a density plot between international research network and overall score.
library(ggplot2)
#Calling our data
data <- top_100_world_university_2024
#Setting the variables
ggplot(data, aes(x = international_research_network, y = overall_score)) +
#Filling our density plot color
geom_density_2d_filled(color = "blue", alpha = 0.3) +
labs(title = "Density Plot",
x = "International Research Network",
y = "Overall Score") +
theme_minimal()
In this density plot between international research and overall score, we are looking to see how much this variable relates to the overall score. As the visual shows, there are several universities that scored in that scored high in overall score but not so high in the international research network and vice versa. From this we can conclude that most likely the international research network may not relate very high to overall score.
Bivariate Boxplot
Our last visual is a bivariate boxplot between international student ratio and international faculty ratio. Bivariate boxplots are used to see the central tendencies of two variables as well as how the variables vary.
#Bivariate Boxplots are in the MVA library so we call it
library(MVA) ## Warning: package 'MVA' was built under R version 4.3.2## Loading required package: HSAUR2## Warning: package 'HSAUR2' was built under R version 4.3.2## Loading required package: tools#Then call our data
data <- top_100_world_university_2024
#Setting our variables
bvbox(data[, c("international_students_ratio", "international_faculty_ratio")], xlab = "Int. Students", ylab = "Int. Faculty")
#Since we want the outliers identified in our plot we set it and make sure they are set to true on the plot
outliers <- identify(data$international_students_ratio, data$international_faculty_ratio, plot = TRUE) 
From this visual, we can see that there are a few outliers within the plot. There seems to be a cluster of both variables near the 100 on the int. students and near the 100 for the int. faculty. There are a few outliers, but from the plot we can gather that most likely from if there is a high international student ratio there is a high international faculty ratio. This may be important for some when picking a school because they themselves may be an international student.
Conclusion
From these findings we can conclude that out of the universities in the data set, the overall score comes from all of the variables. As a college student, when deciding on an institution to attend, there are a variety of things when finally picking one. There could be an addition of variables to the data set. For example, cost could be a variable. Also, some variables can be argued that they are more important than others. For example, employment outcomes could be argued as the most important. When we look at the correlation between academic reputation and employment outcomes, it sits at 53% in this data set. However, it could arguably be one of the most important variables and academic reputation to employment outcomes is one of the selling points of a lot of universities. Overall, the overall score, employment outcomes, international student ratio, international faculty ratio, international research network, and academic reputation were the variables we wanted to focus on and highlight in this data set.I think that this is an informative data set, however, I am not sure I would use this exact data set when choosing a school to attend. There are some variables missing like cost that I think are important factors in the overall score of some of these institutions because at the end of the day a school could be scored as a top university but if it will cost hundreds of thousands of dollars, it may not be feasible for many. Nonetheless, it is a decent and informative data set.