Artifacts of Idiosyncracy in Global Street View Data

2025-05-16

AI Summary

This paper identifies systematic coverage bias in global street-view imagery—arising from inherent urban structural characteristics—where even high-density acquisition yields significant, non-random representational distortions across cities. Method: We quantitatively characterize statistical associations between street-view coverage distribution and urban morphology, road network density, and functional zoning; develop a reproducible framework for coverage bias assessment; and integrate geospatial modeling, multi-city coverage statistics (28 cities), and semi-structured field interviews in Amsterdam to trace bias origins. Contribution/Results: Our analysis reveals that coverage disparities are structurally embedded—not merely sampling artifacts—and systematically correlate with urban form and land-use patterns. The proposed evaluation framework enables rigorous, cross-city bias diagnosis and mitigation. Findings provide a methodological foundation and actionable pathway for constructing fair, representative, and geographically equitable street-view datasets—critical for downstream applications in urban analytics, computer vision, and AI-driven policy.

Technology Category

• Computer Vision: Bias, Fairness & Privacy
• Philosophy and Ethics of AI: Bias, Fairness & Equity
• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy

Application Category

• Social Networks and Social Media: Fairness and bias in social network and social media analysis
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Street view data is increasingly being used in computer vision applications in recent years. Machine learning datasets are collected for these applications using simple sampling techniques. These datasets are assumed to be a systematic representation of cities, especially when densely sampled. Prior works however, show that there are clear gaps in coverage, with certain cities or regions being covered poorly or not at all. Here we demonstrate that a cities' idiosyncracies, such as city layout, may lead to biases in street view data for 28 cities across the globe, even when they are densely covered. We quantitatively uncover biases in the distribution of coverage of street view data and propose a method for evaluation of such distributions to get better insight in idiosyncracies in a cities' coverage. In addition, we perform a case study of Amsterdam with semi-structured interviews, showing how idiosyncracies of the collection process impact representation of cities and regions and allowing us to address biases at their source.

Problem

• Biases in street view data due to city layout idiosyncracies
• Gaps in coverage despite dense sampling in global cities
• Impact of collection process on city representation accuracy

Innovation

• Quantitatively uncover biases in street view coverage
• Propose evaluation method for coverage distribution insights
• Conduct case study with interviews to address biases