Coverage and Bias of Street View Imagery in Mapping the Urban Environment

TL;DR

This paper tackles the representativeness and reliability of Street View Imagery (SVI) in urban studies by introducing an element-level coverage estimation workflow that fuses isovist geometry with semantic segmentation. It demonstrates how SVI coverage on building facades is incomplete and biased, with only 62.4% of buildings visible and an average facade completeness of 12.4%, and finds that non-residential buildings are often over-represented while interior or peripheral neighborhoods are under-represented. The methodology combines geometric visibility, obstruction considerations, and a four-indicator system (CoC-B, FoC-B, CoC-A, FoC-A) to quantify coverage at both building and area scales. A London case study shows that coverage varies with building type, size, and local environmental factors, and that SVI sampling intervals around 50–60 meters balance coverage completeness and redundancy. The work argues for treating element-level coverage as a new data-quality dimension in SVI analyses and provides practical guidance for designing SVI data collection and interpretation in urban research.

Abstract

Street View Imagery (SVI) has emerged as a valuable data form in urban studies, enabling new ways to map and sense urban environments. However, fundamental concerns regarding the representativeness, quality, and reliability of SVI remain underexplored, e.g. to what extent can cities be captured by such data and do data gaps result in bias. This research, positioned at the intersection of spatial data quality and urban analytics, addresses these concerns by proposing a novel and effective method to estimate SVI's element-level coverage in the urban environment. The method integrates the positional relationships between SVI and target elements, as well as the impact of physical obstructions. Expanding the domain of data quality to SVI, we introduce an indicator system that evaluates the extent of coverage, focusing on the completeness and frequency dimensions. Taking London as a case study, three experiments are conducted to identify potential biases in SVI's ability to cover and represent urban environmental elements, using building facades as an example. It is found that despite their high availability along urban road networks, Google Street View covers only 62.4 % of buildings in the case study area. The average facade coverage per building is 12.4 %. SVI tends to over-represent non-residential buildings, thus possibly resulting in biased analyses, and its coverage of environmental elements is position-dependent. The research also highlights the variability of SVI coverage under different data acquisition practices and proposes an optimal sampling interval range of 50-60 m for SVI collection. The findings suggest that while SVI offers valuable insights, it is no panacea - its application in urban research requires careful consideration of data coverage and element-level representativeness to ensure reliable results.

Figures (13)

• Figure 1: Research framework. The isovist analysis example is generated based on 't4gpd' Python library.
• Figure 2: A simplified workflow integrating isovist analysis and computer vision technology for estimating the SVI coverage on building facades.
• Figure 3: The data collection and pre-processing workflow in the case study. Source of the base map data: OpenStreetMap, Greater London Authority.
• Figure 4: Map illustrating the distribution of CoC-B indicator across the case study area.
• Figure 5: Boxplots showing the distribution of CoC-B for different function types and size quantiles.