Linked Data-Driven Decision Support for Accessible Travelling
Chaohai Ding, Mike Wald, Gary Wills · 2015 · Proceedings of the 12th International Web for All Conference (W4A) · doi:10.1145/2745555.2746681
Summary
This doctoral consortium paper proposes using Linked Data principles to build a decision support system (DSS) for accessible travel planning, primarily targeting people with mobility difficulties. The core problem is that accessibility information about physical places — ramps, step-free routes, accessible toilets, transport options — is scattered across isolated systems in incompatible formats, making it expensive and time-consuming for disabled travellers to find what they need. The research reviews existing projects including U-Access (pedestrian routing by ability), OurWay (crowdsourced accessibility annotations), and Accessibility Maps (crowdsourced accessibility mapping), identifying two major problems: a lack of structured accessibility information for physical places on the Web, and data isolation where each project's data cannot be reused by others due to different formats and schemas. The proposed solution has two components: a data integration model that uses Linked Data principles (as defined by Tim Berners-Lee) to publish, integrate, and interlink heterogeneous open accessibility datasets into a unified RDF knowledge base with SPARQL endpoints, and a decision support model that applies optimised algorithms to this knowledge base to recommend accessible routes and travel decisions matched to individual users' capabilities and preferences.
Key findings
The reference architecture comprises several layers. The knowledge and database component stores linked open accessibility data in an RDF store. The data integration model handles data publishing and interlinking, using an ontology-based entity matching approach to resolve duplicates and incomplete records across sources. It connects to external open accessibility data, extracted data, and the broader Linked Open Data Cloud (including LinkedGeoData). The decision support model includes route planning, on-trip assistance, and evaluation components, informed by a user preference model. At the time of writing, the data integration experiments were complete — open accessibility data had been fetched from several systems, interlinked using entity matching algorithms, and published as Linked Data with public SPARQL endpoints. The next phase was optimising DSS algorithms for accessible journey planning based on the generated knowledge base and user capabilities. The architecture also includes a user interface component supporting both human users and machine agents, enabling other applications to consume the accessibility data programmatically.
Relevance
This paper tackles a genuine infrastructure problem for accessible travel: accessibility data exists but is fragmented, inconsistent, and siloed across different projects. The Linked Data approach offers a principled solution for data interoperability — making it possible for different accessible travel tools to share and build on each other's data rather than starting from scratch. For practitioners building accessible wayfinding or travel planning tools, the reference architecture provides a useful template for how to structure data integration from heterogeneous sources. The emphasis on matching recommendations to individual user capabilities (rather than a binary accessible/inaccessible classification) reflects good practice. However, this is a short doctoral proposal (2 pages) presenting architecture and early-stage data integration work without user evaluation or evidence that the DSS algorithms produce better travel recommendations than existing approaches. The practical challenge of keeping accessibility data current — physical conditions change frequently — is not addressed.
Tags: linked data · semantic web · accessible travel · mobility impairment · decision support system · open data · wayfinding · route planning · data integration