EduVision DV: Visual Analytics for University Rankings and Education Indicators
This paper presents EduVision DV, a reproducible data pipeline and interactive Tableau dashboard system that integrates and harmonizes multiple university ranking datasets with World Bank education indicators to enable cross-system comparison while explicitly disclosing underlying data quality limitations such as estimated scores and constant values.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Every year, millions of students, parents, and university leaders turn to global rankings to decide where to study, where to hire, and where to invest. These lists promise a clear hierarchy of academic excellence, sorting thousands of institutions into a single line from best to worst. However, the data behind these lists is far more fragmented than the final headlines suggest. Different organizations measure success in different ways, using different names for the same schools and different formulas to calculate scores. Some numbers are carefully measured, while others are rough estimates or mathematical guesses. When these mismatched pieces are mashed together without a clear map, the resulting picture can be misleading, hiding the fact that a school's high score might be a calculated guess rather than a hard fact.
This is the problem Vaibhav Kumawat set out to solve with a project called EduVision DV. Instead of trying to create a new ranking system or predict which university will win the next year, Kumawat built a tool to make the existing, messy data transparent and comparable. He gathered four massive collections of information: the latest global university rankings from two major providers, a third set of ranking data from the previous year, and a long-term record of national education statistics from the World Bank. These sources originally spoke different languages, with different column names and different lists of schools. Kumawat wrote a computer program to act as a universal translator, cleaning up the names, matching the same universities across all four lists, and stitching them together into a single, coherent database.
The result is an interactive visual guide that allows anyone to explore these connections without needing to be a data scientist. The system links 1,503 universities from the primary ranking list with detailed performance metrics from the other sources. It reveals that while the rankings cover a vast number of institutions, there are significant gaps. For instance, only about half of the universities in the main list appear in the other two ranking datasets, meaning a direct comparison is not possible for many schools. More importantly, the tool exposes the hidden nature of the scores themselves. In the primary ranking list, more than 60 percent of the overall scores for universities ranked 601 and below are not actual measured values but are estimated by the ranking organization. Similarly, in the other datasets, large portions of the data are either reconstructed using formulas or are constant values applied to groups of schools that were not individually ranked.
By making these distinctions visible, the project changes how the data can be read. When a user looks at an average score across all universities, the tool shows that this number is a mix of real measurements and estimates, which can skew the perception of overall performance. The analysis also found that a university's standing is heavily influenced by its reputation and research citations, with these factors showing the strongest link to the final ranking. However, the tool also highlights that national wealth and government spending on education, while important, are not the only drivers of success, and that the data from national statistics often covers a different time period than the current rankings, making direct cause-and-effect conclusions impossible.
The final product is a set of five connected screens that let users filter by region, country, or specific metrics like student-to-teacher ratios or international student numbers. It does not offer a definitive answer to which university is the best, but rather provides a clear view of the landscape, showing where the data is solid and where it is an approximation. This approach respects the complexity of the information, ensuring that users do not mistake a calculated estimate for a measured reality. In a field often dominated by simple lists and headlines, this work offers a more honest way to understand the true state of global higher education, reminding us that behind every rank is a specific set of data with its own limits and stories.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.