Accessibility maps attempt to communicate spatial attributes of the built environment so that pedestrians with diverse sets of mobility requirements can plan accessible trips and planners and other decision-makers can make informed choices about infrastructure improvements. Both of these tasks have the underlying path-based (and therefore, network-representable) goal of traversing the environment from one place to another. However, extant geodata for pedestrian infrastructure has traditionally been collected as disconnected assets, resulting in ambiguous accessibility maps that overlay inaccurate, incomplete, non-standardized, and/or disconnected geometries on top of a base map. As a result, pedestrian networks are typically unsupported by proprietary, open source, or municipal data. In this talk, I compare the informational content of different mapping approaches, discuss user experiences and evaluations of a subset of these approaches, and present two new accessibility tools based on a true pedestrian network: AccessMap, a public-facing pedestrian trip planning map, and SidewalkScore, a network-based metric of sidewalk accessibility. Findings: Study participants primarily choose network-based trip planning approaches, even when those approaches do not describe the pedestrian infrastructure of great importance to them, and down-rate asset-based accessibility maps when exposed to network-based trip planning services.