A data analytics approach for university competitiveness: the QS world university rankings

Abstract

AbstractIn recent years, higher education has felt pressured to prepare its graduates for the highly competitive international market due to globalization. Thus, many institutions have turned to position themselves well in university rankings as a way to attract the best academic and student talent from all over the world. Our work presents a predictive model for measuring university performance in the QS world university rankings (QS-WUR). We used a ten-year dataset to build models with statistical and machine learning algorithms contained in the library Caret of the RStudio software tool, to forecast global university position in QS-WUR. With these tools, we designed a methodology to predict the university partners' Final Scores based on their historical performance, achieving errors in the range of one or two points out of 100. The modelling may be a useful aid for university officers to develop strategies for improving institutional processes to attract the best students, faculty, and funding, enhance international collaboration and outlook, and foster international university prestige.