Integrating Power BI and PLS-SEM dashboards to model vocational expectations, tutorial quality, and academic problems in higher education

–This study presents a structural model linking career expectations, tutorial quality, and academic difficulties among undergraduate engineering students, enhanced by developing real-time educational dashboards with Power BI. The primary objective was to evaluate whether aligning students' career aspirations with tutorial experiences could improve perceived tutoring quality and contribute to the early detection of academic difficulties. A sample of 186 Systems Engineering students from the National University of Central Peru provided the data for this study. Using Partial Least Squares Structural Equation Modeling (PLS-SEM), the model showed high reliability (Cronbach's alpha: 0.717 to 0.882), convergent validity (AVE: 0.679 to 0.773), and discriminant validity (HTMT < 0.90). The results indicated a strong positive influence of vocational expectation on tutoring quality (β = 0.782; R² = 0.612), while tutoring quality showed a weak negative relationship with academic problems (β = –0.052; R² = 0.003). Predictive accuracy reached 82%, and statistical assumptions were validated using the Cramér-von Mises test (p < 0.001). To improve interpretation and decision-making, interactive dashboards were developed in Power BI, including radar charts, heat maps, and risk matrices. These visual tools allowed academic advisors to identify patterns across semesters, customize tutoring strategies, and monitor vocational alignment over time. Despite the limited explanatory power regarding academic performance, the integration of structural modeling and data visualization offers a scalable framework for evidence-based tutoring in higher education. Future studies are encouraged to incorporate emotional, motivational, and metacognitive constructs to strengthen prediction and intervention models.