Future workspace needs flexibility and diversity: A machine learning-driven behavioural analysis of co-working space

Area of Science:

• Architecture and Urban Planning
• Human-Computer Interaction
• Occupational Psychology

Background:

• The evolution of workspaces is driven by technology, shifting work patterns, and employee well-being.
• Co-working spaces offer cost-effectiveness, flexible design, and multi-use opportunities as an alternative workspace solution.
• Understanding user behavior and space utilization is crucial for optimizing modern office environments.

Purpose of the Study:

• To examine human-centric design choices in a flexible co-working space in London.
• To investigate spatial and temporal variations in occupancy levels and user behavior.
• To provide data-driven insights for future office space planning, particularly for hybrid working.

Main Methods:

• Machine-learning-driven analysis of sensor-detected occupancy data (145 million data points over 477 days).
• On-site observations of occupant activities, environment, behavior, and preferences (1000+ snapshots over 13 days).
• Decomposition of space usage, calculation of seat utilization, and identification of spatial hotspots.

Main Results:

• Shared working areas near windows or in open, visible locations are preferred for communication and general working.
• Semi-enclosed spaces with higher privacy are favored for focused, individual work.
• The flexibility of multi-use opportunities was identified as the most preferred feature for hybrid working.

Conclusions:

• Human-centric design significantly impacts the functionality and user satisfaction of co-working spaces.
• Data-driven insights into space utilization and user preferences are essential for effective office planning.
• Findings support the design of future workspaces that accommodate hybrid working, hot-desking, and co-working models.