運用多維度分析及資料視覺化於校務研究之教師研究産能分析

隨著高等教育規模快速擴張與全球市場競爭的壓力，許多國家投入高等教育改革，以提升經濟生產力及國際競爭力為目標，以知識經濟為主軸邁進，世界大學排名成為各大學追求的目標，以最受矚目的三個排名系統為例，其評比教師在研究產能表現之權重相對高。本研究旨在運用Power BI資料視覺化探討個案大學105-109年共5年的專任教師資料，從教師人口變項、研究領域、教師職級、是否擔任行政職等構面來觀察教師研究產能是否有差別，並運用SPSS Modeler決策樹建構出教師研究產能的預測模型。研究結果發現，45歲以下的副教授或正教授，在研究的量及質上的表現都具有潛力；工程學院、電機通訊學院46-55歲正教授注重研究的質；56歲以上正教授注重研究的量，且構面影響重要度最高的是教師年齡及職級，而教師是否擔任行政職並不會造成太大的影響。決策者可透過視覺化儀表盤點教師研究產能的分布狀況及研究品質，盤點整體研發能量，輔以預測模型訂定合宜的獎勵制度及教師聘任的汰除機制，達到解釋、預測、控制的科學目的，最後，提出學校政策面、資訊系統面、教師研究面之建議，以期未來更精準呈現教師研究產能預測模型的完整。

關鍵字

高等教育改革；校務研究；教師研究產能；資料視覺化；決策樹

並列摘要

With the rapid expansion of higher education and the pressure of competition on the global market, many countries have carried out higher education reform with the aim of increasing economic productivity and international competitiveness, with the knowledge economy as the main axis and the world university rankings have become a major objective for universities. The goal is to take the three most prominent ranking systems as examples. The weighting of teachers' research productivity is relatively high. The purpose of this research we use Power BI to visually explore the data of full-time teachers in a case university from 105 to 109. From the dimensions of teacher demographic variables, research fields, teacher rank, and whether they are administratively responsible to observe whether teachers' research productivity are differences and we will use SPSS Modeler decision trees to construct a prediction model.The results of the study, it emerged that associate professors or full professors under the age of 45 have the potential in both quantitatively and qualitatively performance; the full professors aged 46-55 of Engineering and Electrical and Communications focus on the quality of research; full professors above 56 focus on the quantity of research, and the most important factors are the age and rank of teachers, and whether they hold administrative positions are less important. Decision makers can use the visual instrument to distribution of teachers' research capacity and research quality, as well as to determine the overall research and development energy. Supplemented by the prediction model to formulate a suitable reward system and an elimination mechanism for teacher appointment can be developed to achieve scientific purpose of interpretation, prediction, and control. The report concludes by offering proposals on school policy, information systems, and teacher research, in order to improve the accuracy of the teacher development prediction model in the future.