Delineating potential DRT operating areas: An origin–destination clustering approach

Abstract

Investment in Demand-Responsive Transport (DRT) has emerged as a sustainable transport intervention option for areas that are traditionally hard to serve by high frequency public transport. When used as a first- and last-mile feeder, DRT has the potential to reduce car dependency and enhance access to the wider network. However, many DRT schemes fail—often due to overly flexible, poorly targeted service areas that do not align with actual travel patterns, making efficient pooling difficult. While planners may already have a general sense of where DRT might be useful, there is limited guidance on how to identify precise operating zones based on spatiotemporal demand. This paper presents a method for identifying potential DRT service areas using spatial clustering of origin–destination (OD) flows. We apply the method in Leeds, UK, focusing on OD pairs with poor public transport supply and low potential demand. The approach identifies spatial clusters where demand is both underserved and sufficiently concentrated to support DRT operation. By narrowing service areas to zones where pooling is more likely and where DRT complements rather than competes with fixed-route services, the method helps address two key challenges in DRT planning. The results offer a reproducible, data-driven input for delineating preliminary DRT service areas—supporting strategic planning, integration with downstream agent-based models, and further refinement through local knowledge. The method provides a foundation for future work on designing DRT services that complement the public transport network, particularly in low-density urban peripheries. • New method to delineate DRT areas based on travel demand and PT coverage. • OD clustering algorithm identifies clusters of unmet travel demand. • Leeds case study shows DRT zones linking suburbs to core transit. • Open, reproducible code allows replication and extension of this work.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 5 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.