Graph-Based Optimisation of Network Expansion in a Dockless Bike Sharing System

TL;DR

The paper tackles dockless bike-sharing expansion by constructing an optimized geo-temporal graph from Moby Bikes trip data and applying hierarchical agglomerative clustering to reduce spatial complexity, then using the Louvain algorithm on multi-granularity networks $G_{Basic}$, $G_{Day}$, and $G_{Hour}$ to detect communities and validate candidate stations. It introduces a novel network-expansion algorithm with explicit clustering and distance thresholds, and demonstrates that 146 new stations (to total 238) can be added while preserving comparable usage patterns to existing stations, with strong self-contained communities indicated by modularity values up to $Q=0.54$ at the hourly granularity. The work provides data-driven guidance for station placement and fleet redistribution, showing how spatiotemporal patterns influence expansion decisions and operational efficiency. Practical impact includes improved redistribution efficiency and targeted expansion strategies for dockless BSS operators, supported by a scalable graph-based methodology and multi-granularity analysis.

Abstract

Bike-sharing systems (BSSs) are deployed in over a thousand cities worldwide and play an important role in many urban transportation systems. BSSs alleviate congestion, reduce pollution and promote physical exercise. It is essential to explore the spatiotemporal patterns of bike-sharing demand, as well as the factors that influence these patterns, in order to optimise system operational efficiency. In this study, an optimised geo-temporal graph is constructed using trip data from Moby Bikes, a dockless BSS operator. The process of optimising the graph unveiled prime locations for erecting new stations during future expansions of the BSS. The Louvain algorithm, a community detection technique, is employed to uncover usage patterns at different levels of temporal granularity. The community detection results reveal largely self-contained sub-networks that exhibit similar usage patterns at their respective levels of temporal granularity. Overall, this study reinforces that BSSs are intrinsically spatiotemporal systems, with community presence driven by spatiotemporal dynamics. These findings may aid operators in improving redistribution efficiency.

Figures (7)

• Figure 1: The candidate graph generated by HAC, including the pre-existing stations. Nodes are shown in purple and edges in yellow.
• Figure 2: The selected graph includes pre-existing stations and selected new stations. Node size is scaled according to the number of self-contained trips (self-edges). Edge width is scaled according to the number of directed trips between nodes (out-edges). Only edges with a weight in the top 1% percentile of weights are shown.
• Figure 3: Community detection for $G_{Basic}$. Stations are coloured according to their respective community assignment.
• Figure 4: Community detection for $G_{Day}$.
• Figure 5: Daily travel patterns per community in $G_{Day}$