Data Driven Approach to Forecast Building Occupant Complaints

Occupant complaints are a reflection of poor building performance and an unsatisfactory indoor environment. One way to mitigate those complaints and to ensure occupants' satisfaction with regards to building performance is through a well-performing facility management that is capable of planning for and addressing maintenance and repair (M&R) services. This paper proposes a data analytics and machine learning framework to predict and analyze occupants' complaint data as part of the facility management's predictive maintenance approach. The framework was tested for a period of one year on a highly unstructured and unsolicited occupants' complaints data recorded by facility management operators in a residential complex. Text mining results showed that air conditioner (AC) complaints are among the most frequent complaints in the dataset under study. Using the built nonlinear autoregressive exogenous (NARX) prediction model to forecast such complaints resulted in acceptable validation and test mean square errors of 0.822 and 0.188 respectively. Ongoing works aim at expanding this framework to include a larger data set and to develop a staffing plan used to handle those complaints thus enhancing occupant satisfaction and building performance.