The Asymmetry of Prioritization: Comparing Expert-Informed and Data-Driven Weighting for a Bike Equity Index in the San Francisco Bay Area, U.S.

The distribution of cycling infrastructure across a region depends not only on planning priorities but also on how equity is measured. Composite bike equity indices are increasingly used to guide infrastructure investment, yet the weighting schemes embedded in these indices can substantially influence which communities are identified as underserved. Despite their growing use, little empirical evidence exists on whether expert-informed and data-driven weighting approaches produce comparable equity assessments. This study addresses that gap by comparing two weighting methods applied to the same validated Bike Equity Index across 4508 census block groups in the nine-county San Francisco Bay Area: an expert-informed Delphi Budget Allocation approach (Delphi-BAL) and a data-driven Principal Component Analysis approach weighted by explained variance ratios (PCA-EVR). Supply and demand indices were compared using paired t-tests, Pearson and Spearman correlation analyses, and Global Moran’s I. On the demand side, the two approaches showed near-perfect agreement (r = 0.984) and a small practical difference (Cohen’s d = −0.59), indicating that they identify similar high-need communities. On the supply side, however, the methods diverged substantially, with a large effect size (d = −2.04), moderate correlation (r = 0.797), and significant spatial clustering of score differences (Moran’s I = 0.372). These findings suggest that agreement between expert-informed and data-driven weighting depends on the equity dimension being assessed. While data-driven methods may provide results comparable to expert-informed approaches for some applications, they may produce different prioritization patterns where statistical covariance does not align with normative planning priorities. The results highlight the importance of combining data-driven analysis with expert and community validation to support equitable, context-sensitive infrastructure investment decisions.