Faculty performance evaluation is the conventional method used among Higher Education Institutions (HEI) to gain needed insights as to the basis for important decisions that may involve faculty tenure, promotion, and scholarships. Most often, these evaluations utilized numerical ratings to measure the faculty performance while textual feedback may have no bearing at all. As such, textual feedback has been widely adopted to measure faculty performance through text mining. However, as student's textual comments or feedback are subjective in nature, they may not capture all insights that are normally found in a standard survey questionnaire. Thus, faculty performance evaluation, which are solely based on text comments may not in turn give reliable results. Hence, the need to combine numerical ratings and textual feedback as a measure of the overall faculty performance. In this study, we introduce an innovative approach in combining numerical ratings and textual feedback in measuring the overall faculty performance. We leverage the use of Fine-Grained Sentiment Analysis to classify the textual feedback into five sentiment scores. We then assigned a numerical weight score to each polarity to calculate the overall faculty performance combined with the numerical ratings. We developed a web-based faculty evaluation system to streamline the collection of both the numerical and textual evaluation data using the instrument used by all State Universities and Colleges in the Philippines (i.e., QCE NBC 461) integrating the sentiment analysis module created. The system was deployed in University of Science and Technology of Southern Philippines. Experiments reveal that our proposed approach is efficient and provides effective results in terms of the system's usability and functionality.