Analysing and visualising bike-sharing demand with outliers

Nicola Rennie,Catherine Cleophas,Adam M. Sykulski,Florian Dost

DOI: https://doi.org/10.48550/arXiv.2204.06112

2023-01-31

Abstract:Bike-sharing is a popular component of sustainable urban mobility. It requires anticipatory planning, e.g. of station locations and inventory, to balance expected demand and capacity. However, external factors such as extreme weather or glitches in public transport, can cause demand to deviate from baseline levels. Identifying such outliers keeps historic data reliable and improves forecasts. In this paper we show how outliers can be identified by clustering stations and applying a functional depth analysis. We apply our analysis techniques to the Washington D.C. Capital Bikeshare data set as the running example throughout the paper, but our methodology is general by design. Furthermore, we offer an array of meaningful visualisations to communicate findings and highlight patterns in demand. Last but not least, we formulate managerial recommendations on how to use both the demand forecast and the identified outliers in the bike-sharing planning process.

Applications

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the identification and analysis of demand outliers in the bike - sharing system. Specifically, the paper focuses on how to identify those situations where the demand deviates from the normal level due to external factors (such as extreme weather or public transport failures) through clustering stations and applying in - depth functional analysis. These outliers will affect the reliability of historical data and, in turn, the accuracy of future predictions. Therefore, identifying and dealing with these outliers is crucial for improving the planning efficiency and service quality of the bike - sharing system.

The main contributions of the paper include:
1. Conducted an in - depth analysis of the time - use patterns of the Capital Bikeshare service.
2. Proposed a spatial clustering method for bike - sharing stations based on geographical proximity and similarity of use patterns.
3. Investigated the time trends of the detected outliers and their possible causes.
4. Analyzed the spatial distribution patterns of the detected outliers.

Through these methods, the paper aims to provide a general and data - driven methodology that can be applied not only to the Capital Bikeshare dataset in Washington D.C., but also extended to other bike - sharing datasets around the world, thus helping relevant enterprises and managers to carry out demand forecasting and resource allocation more effectively.