A Data-Driven Customer Satisfaction Survey Model Integrating Sentiment and Prescriptive Analytics in Higher Educations Institutions

Abstract

This research presents the development and implementation of A Data-Driven Customer Satisfaction Survey Model Integrating Sentiment and Prescriptive Analytics in Higher Educations Institutions, integrating sentiment analysis and prescriptive analytics to strengthen institutional decision-making. Traditional survey instruments at PSU have relied mainly on numerical ratings, which often fail to capture the deeper meaning behind stakeholder feedback. This study addresses that gap by using advanced data analytics to gain a clearer, more nuanced understanding of stakeholder experiences and to provide actionable recommendations for continuous improvement. The primary goal of the research was to design, validate, and deploy a survey system capable of systematically gathering and analyzing both quantitative and qualitative feedback from students, faculty, staff, and other stakeholders across PSU campuses. A mixed-method approach was used, combining structured survey questions with open-ended responses. The textual feedback gathered from these comments underwent several preprocessing steps, including text cleaning, tokenization, stop-word removal, lemmatization, and vectorization using BERT embeddings. These steps allowed the study to classify sentiments accurately as positive, negative, or neutral, offering insights that numbers alone could not provide. To make the findings more useful for decision-makers, the study incorporated a prescriptive analytics framework. This framework transformed sentiment results into strategic, prioritized recommendations for university leaders. The process involved data collection, pre-processing, feature extraction, model selection, training, and validation. Machine learning and artificial intelligence tools were used to ensure high accuracy in sentiment classification, improving the reliability of feedback interpretation. The prescriptive analytics outputs provided PSU administrators with clear action points to reinforce institutional strengths and address areas needing improvement. The findings revealed that, overall, stakeholders were satisfied with PSU’s academic services, facilities, and administrative support. However, several recurring concerns emerged. These included issues related to technological infrastructure, communication efficiency, and student support services. The use of sentiment analysis helped uncover patterns of dissatisfaction that traditional surveys often overlook. The integration of prescriptive analytics further translated these insights into specific recommendations, such as upgrading digital platforms for student services, enhancing faculty development initiatives, and improving feedback and communication mechanisms. This study contributes significantly to the field of educational management by demonstrating the practical use of artificial intelligence—specifically sentiment analysis and prescriptive analytics— in understanding and responding to stakeholder needs. It offers a replicable framework for other higher education institutions seeking to modernize their feedback systems and adopt data-driven decision-making practices. The research supports the shift toward evidence-based governance and reinforces the importance of continuous quality assurance in academic environments. In conclusion, the newly developed survey tool strengthened PSU’s capacity to capture meaningful feedback and translate it into actionable institutional strategies. By revealing both strengths and hidden challenges, the tool promotes a more responsive, transparent, and accountable governance system. The study recommends integrating this tool permanently into PSU’s quality assurance processes and encourages further research to explore its use in other universities and organizational settings.