Capturing usage patterns in bike sharing system via multilayer network fused Lasso

TL;DR

This paper introduces a multilayer network fused Lasso method to analyze complex spatio-temporal bike-sharing data, capturing usage patterns without arbitrary data partitioning, and demonstrating competitive predictive performance.

Contribution

It proposes a novel penalized regression approach with multilayer network fused Lasso to model spatio-temporal features in bike-sharing data, accounting for station-specific behaviors.

Findings

Achieves competitive predictive accuracy on real datasets

Provides new insights into bike usage patterns

Effectively captures complex spatio-temporal dependencies

Abstract

Data collected from a bike-sharing system exhibit complex temporal and spatial features. We analyze shared-bike usage data collected in three large cities at the level of individual stations, accounting for station-specific behavior and covariate effects. For this, we adopt a penalized regression approach with a multilayer network fused Lasso penalty. These fusion penalties are imposed on networks which embed spatio-temporal linkages, and capture the homogeneity in bike usage that is attributed to intricate spatio-temporal features without arbitrarily partitioning the data. On the real-life datasets, we demonstrate that the proposed approach yields competitive predictive performance and provides a new interpretation of the data.