Assessing automation: Methodological insights from experimenting with computer vision for public life research. Issue 3 (1st September 2020)

• Record Type: Journal Article
• Authors: Schlickman, Emily
• Is Part Of: Journal of landscape architecture. Volume 15:Issue 3(2020)
• Page Start: 48
• Page End: 59
• Publication Date: 2020-09-01
• Subjects: spatial data collection -- research methodology -- automation -- computer vision -- public life; Landscape architecture -- Periodicals; Architecture du paysage -- Périodiques; Landscape architecture; Periodicals; Electronic journals; 712.05
• DOI: 10.1080/18626033.2020.1886515
• Languages: English
• ISSNs: 1862-6033

Abstract

Abstract: Today, there is a growing interest in designer-led assessment efforts focused on social performance. While this interest has led to significant developments in our understanding of how people occupy urban public space, the spatial data-collection methodologies that are currently being employed are similar to those used in early foundational studies. This paper explores how advancements in digital automation might expand the designer toolbox and, in turn, help to inform design decision making. To do this, the paper unpacks a recent research project that experimented with an automated tabulation technique using computer vision to visualize the spatial distribution of urban public space users. The goal of the computer-vision output is to help designers observe and identify common social patterns across ten plazas in New York City. A key finding from the study is that the automated data-collection technique employed does not fully replace manual techniques, but is complementary.