Analysis of building maintenance requests using a text mining approach: building services evaluation

Abstract

End users’ maintenance requests gathered from computerized maintenance management systems (CMMS) configure a rich source of information to evaluate the occupants’ satisfaction and the building systems. Nevertheless, the non-standardized data gathered from CMMS makes it difficult to carry out analytics. This paper presents a text mining approach to extract information from end users’ maintenance requests and an analysis of 6,830 maintenance requests derived from 46 buildings including offices, academic buildings and laboratories over two and a half years. The research results reveal that the most common maintenance requests during building operation and maintenance are related to problems in electrical and HVAC fixtures. Although the year of construction is not related to the occupants’ maintenance requests, the type of building use and building property do influence them. The implementation of control and preventive strategies based on these results may increase facility managers’ productivity and building systems’ performance.