Planned Preventive Maintenance (PPM) and Unplanned Maintenance (UPM) are the most common types of facility maintenance. This paper analyzes current trends and status of Facility Management (FM) practice at higher education institutions by proposing a systematic data-driven methodology using Natural Language Process (NLP) approaches, statistical analysis, risk-profile analysis, and outlier analysis. This study utilizes a descriptive database entitled “Facility Management Unified Classification Database (FMUCD)” to conduct the systematic data-driven analyses. The 5-year data from 2015 to 2019 was collected from eight universities in North America. A preprocessing step included but was not limited to identifying common data attributes, cleaning noisy data, and removing unnecessary data. The outcomes of this study can facilitate the decision-making process by providing an understanding of various aspects of educational facility management trends and risks. The methodology developed gives decision makers of higher education institutions, including facility managers and institution administrators, effective strategies to establish long-term budgetary goals, which will lead to the enhancement of the asset value of the institutions.
Historical building maintenance data were collected from twelve universities in the United States (U.S.) and Canada, spanning from the year 2002 to 2021, with a particular focus on Planned Preventive Maintenance (PPM) and Unplanned Maintenance (UPM). The collected data underwent preprocessing and organization based on the Facility Management Unified Classification Code (FMUCO), which was developed to classify heterogeneous building maintenance data. The resulting dataset comprises nine attribute groups (university, building, system, subsystem, component, work order, work order cost, work order labor, and weather) and their corresponding data attributes. The dataset aims to (1) provide insights into the current status of building management for campus-sized institutions and (2) facilitate data-driven analyses of facility management.
This research presents a systematic approach for assessing component interdependencies within building subsystems in relation to unplanned maintenance (UPM). Utilizing a three-step methodology—quantitative content analysis, component causal relationships, and interdependency analysis—the study leverages work order description data to identify key components and their interactions. By applying social network analysis to top keyword and term frequency data, the study examines component centrality, revealing critical interdependencies that impact maintenance needs. Findings show that specific components, such as the steel damper and supply fan, play prominent roles within the subsystem network, with significant ripple effects in case of failure. This methodology not only provides valuable insights for facility managers in prioritizing maintenance efforts but also offers a foundation for future predictive maintenance strategies.