Skip to main content

Jing Huei Huang

CGA Postdoctoral Research Fellow


Date: 10/01/22 - 12/01/23
Amount: $89,763.00
Funding Agencies: KOMPAN, Inc.

Play is essential to developing physical and cognitive health for children. Children’s free play appears to be motivated by play environments, such as playgrounds and outdoor recreational settings. Understanding behaviors in playspaces (e.g., where; how using specific spaces and equipment; duration; and variations across groups) will improve design, layout, programming, and management to encourage diversity of play and hopefully a lifelong love and enjoyment of play and the outdoors across our diverse communities. It is currently challenging to analyze play patterns as children’s free play is spontaneous, creative, interactive, and changes over time and across spaces. Traditionally, play episodes were observed and annotated through behavior mapping based on activity types and start and end times of entering a designated play area (Luchs & Fikus, 2013; Sumiya & Nonaka, 2021). This approach is labor-intensive, records activities in play areas designated by researchers, and is often limited to one observation at a time. Wearable and quantitative approaches have been adopted to investigate children’s play patterns using accelerometers and global positioning system (GPS) technology. Our team has started to spatially-aggregate activity points overlaid with distinct playground playspaces (e.g., swing bay; slide) to demonstrate how children’s activity differs across these areas. We have also conducted a hotspot analysis to identify play areas where children tend to be more physically active (i.e., clustering of high intensity activity) compared to less physically active (i.e., clustering of low intensity, or no activity), however this method excludes duration from the clustering analysis. These initial data efforts have revealed a set of exploratory questions and aims, such as duration of play per space, individual, family, and group differences (e.g., sibling pair; gender differences), and how patterns change across time of day, week, season, and across age groups. With recent updates to density-based clustering methods, which identifies groups of points that cluster together in space and time, we have the opportunity to systematically identify play episodes through an unsupervised machine learning approach. This approach could provide valuable information for practitioners by identifying and mapping natural play patterns, better characterizing playspace and amenity use, and correlating play episodes and potential with specific structures, natural elements, social aspects, and layouts of playspaces.

View all grants