A data-driven analysis of occupant workspace dissatisfaction

Studies often aim to determine which indoor environmental quality parameters best predict the overall workspace assessment. However, this method overlooks important differences distinguishing satisfied and dissatisfied occupant groups. We used a new analytical approach on 36671 post-occupancy evaluation responses overcome problem better understand satisfaction in office buildings. Principal components analysis reduced votes with 15 different IEQ items into two principal related to: 1) privacy amount of space, 2) cleanliness maintenance. grouped data by occupants that were either or their workspace. component 1 explained half variability dataset reliably distinguished from those dissatisfied. support vector machine classify groups based 2. Classification was highly accurate (approximately 90%) predominantly space. Further analyses showed generally all other items. There greater independence between attributes for Issues available space an overwhelming determinant dissatisfaction irrespective success attributes. These findings suggest efforts improve workspaces should leverage designs ensure provide sufficient work.