A Data-Driven Framework for Student Satisfaction: Novel Hybridization of Clustering and Performance Mapping Analysis

Service quality can be measured by measuring customer satisfaction with the services provided by an institution or organization. The higher education institutions that provide educational services to students also applied the service quality satisfaction analysis. This study proposes a combined method combining machine learning through the clustering method and the importance-performance analysis or quadrant analysis approach. Combining these two approaches is intended to produce groups of students with varying satisfaction levels. Furthermore, each group will be explored more deeply by using the quadrant method to determine which aspects should be prioritized by the institution to improve its services. The data processing results from 34,087 respondents obtained three groups of students, with the characteristics of the first group having a very satisfied perspective of 36%, the second group having a satisfied perspective of 57%, and the third group having a somewhat satisfied perspective of 7%. The indicators whose services should be improved according to the first group are teaching method (P2), teaching timeliness (P7), and information system for academic services (P21). Whereas for the second group, no service priority was found on indicators that needed improvement. Meanwhile, in the third group, the service indicators lecturer openness (P10), transparency in grading (P15), and friendliness of staff service (P16) were found to be prioritized by the university for improvement.