Sustainable Building Performance through Improved Building-Occupant Interactions: An integrated Data Collection, Modeling and Analysis Approach

  • Min Lin

Student thesis: Doctoral Thesis


There is a growing interest to study, model, and understand the interaction between buildings and their users. On the one hand, actions taken by occupants and facility managers can have tremendous positive – or negative – impact on the levels of energy consumed. On the other, the indoor environmental conditions of buildings can significantly affect occupants' comfort, wellbeing, health, happiness levels, and productivity. Various research efforts exist on individual metrics of building performance; however, studies that integrate these elements are still very scarce in the literature. Consequently, researchers are facing important limitations to study and optimize the performance of the built environment holistically, while accounting for the 'human' dimension of the problem. This dissertation presents an integrated approach to evaluate and improve the interaction between people and their built environment. The proposed framework is multidisciplinary and combines methods such as machine learning, human or 'agent-based' modeling, data collection, and statistical modeling. The capabilities of the framework are demonstrated through case studies using local and regional data, leading to important findings. These include but are not limited to: (1) demonstrating the influence of physical characteristics of buildings on energy consumption and the need for systematic benchmarking of performance; (2) quantifying energy consumption drivers of building occupants in different environments while observed significant spillover effects; (3) exploring multi-dimensional drivers of occupants' perceptions of their built environment and demonstrating their threshold-based relationships; and II (4) confirming the role of occupants' demographical characteristics on predictive models of comfort and productivity, confirming the need to better account for them when assessing building performance. Beyond the findings and recommendations from the case studies, the proposed methods are scalable and illustrate the multidisciplinary approach that is needed to address the current and future multifaceted sustainability challenges facing our built environment.
Date of AwardMay 2020
Original languageAmerican English
SupervisorElie Azar (Supervisor)


  • Occupant Behavior
  • Building-Human Interactions
  • Building Performance
  • Machine Learning
  • Indoor Environmental Quality (IEQ).

Cite this