World University Rankings Analysis

PROBLEM:

Every year, universities throughout the world are rated and ranked, allowing students to pick the best institutions for their future. These different rankings systems are widely used but can be controversial due to potential biases. How can students pick the best universities for them?

GOAL:

This project was conducted to analyze three different ranking systems and aimed to find the best universities and locations in each system. The rankings were also compared against each other to look for any major differences.

SKILLS:

Data Wrangling, Cleaning, and Subsetting
Combining and Exporting Dataframes
Geographic Visualizations
Regression Analysis
Cluster Analysis
Time Series Analysis
Data Visualization and Storytelling

TOOLS:

Python
Tableau
Excel
Powerpoint

DATA:

KEYNOTES:

Due to the large amount of data, subsets were created for each ranking system. Only the top 200 universities were considered for the Center for World University Rankings (CWUR) and Times Higher Education World University Rankings (Times) datasets whereas only the top 100 universities were considered for the Academic Rankings of World Universities (Shanghai) dataset.
Each ranking system factored in different criteria when evaluating and ranking universities. While some criteria were similar in each system (ex: quality of faculty/education), other criteria were present in only one system (ex: income score, male-to-female ratio, etc).
The three ranking systems contained data spanning different years. The Shanghai Rankings had data from 2005 to 2015 whereas the CWUR and Times Rankings only had 2012-2015 and 2011-2016 data, respectively.

World University Rankings Key Insights

Which universities are the best of the best?

The top 5 universities in each ranking systems were found and summarized in a table.
All three systems had Harvard, Stanford, and MIT within their top 5.
Other universities varied between rankings such as Oxford, Cambridge, and UC Berkeley.

Where are the top universities located?

Geographic maps were generated for all three datasets based on university counts and the results were consolidated in a pie chart.
For all three ranking systems, the USA has the most amount of top universities by far.
Other countries with large amounts of universities varied between ranking systems. Some of these include the United Kingdom, France, Germany, and Canada.

Do higher numbers of top universities correlate to higher general rankings of countries?

Linear regression was conducted on each dataset to see if a country's average university ranking was correlated to the number of top universities.
While the variables seemed to be generally correlated, linear regression was deemed to be an inadequate fit due to a low R-squared value in each dataset.

How do rankings change over time?

Time series analysis via forecasting was conducted to see how the top universities would change over time in each dataset.
Most of the top 5 universities in each dataset remained within the top 5 in the forecasted years and were at worst in the top 10.

Recommendations

Top Universities/Countries

Students should heavily consider institutions in the United States since it had the largest number of top universities in each system.
- Harvard, Stanford, and MIT are all great choices when using any of the three ranking systems.
- While some other countries have better average rankings in some years, the sheer number of top tier universities make the USA the best choice .

Future Analysis

Updated ranking information for each dataset would be useful in predicting top universities in more modern times since the most up-to-date information is from 2016.
- Stronger tools than Tableau's forecasting function could possibly predict more accurate values.
Finding missing information for certain criteria would lead to more accurate analysis, as some variables had to be omitted from analysis due to large amounts of blank data.

Final Deliverables

Page updated

Google Sites

Report abuse