Investigating the potential, opportunities and challenges of using crowd-sourced geographic data to support decision making in transport planning

作为一门学科，交通规划在很大程度上依赖于数据洞察来为决策提供依据。由于这些决定的重要性和高影响力，洞察力来自“权威”或“可靠”的数据集，如流量统计、家庭调查和最近的移动电话数据。这些方法通常昂贵、耗时且缺乏空间/时间覆盖(Jestico等人)。2016)。众包是一种使用个人/传感器的参与式数据收集方法，可用于构建数据集。假设众包数据有可能在交通规划中产生新的见解，特别是对交通和日常交通模式的看法。然而，有必要全面审查使用众包数据的挑战和偏见，以确保其符合目的。有鉴于此，本研究旨在回答以下研究问题：-与传统数据源相比，众包数据对移动性有何理解？-谁对众包数据集做出了贡献，他们的动机如何影响数据质量和空间/时间偏差？-在移动性背景下，众包地理数据面临哪些挑战和偏差？如何减轻偏差并确保数据适合使用？-如何将众包数据可靠地应用于交通建模？研究问题将通过研究众包数据的偏差及其来源来实现。将设计一个原始框架来评估众包数据质量并减轻发现的偏差。最后，还将进行一项案例研究，在以代理人为基础的运输模型中实施众包数据，以确定关于如何在运输建模中适当应用新的众包数据集的新指南。这项研究提供了一个关键的机会来挑战交通模型中被认为是‘日常数据’的东西。交通提案有一个关键的机会，可以通过众包数据来丰富，从而能够更大地推动“数据驱动的决策”，这是ESPRC的一个研究优先事项，特别是在信息和通信技术主题方面。这项研究将考虑众包传感器数据的可扩展性、互操作性和可靠性方面的挑战，这是EPSRC的ICT网络和分布式系统领域的关键战略重点。有人说，研究人员应该将研究与现实世界的测试联系起来，这项研究将把数据应用于交通模型，以突出其适用性和现实世界的社会效益。

As a discipline, transport planning is heavily reliant on data insights to inform decisions. Due to the importance and high impact of such decisions, insights are derived from "authoritative" or "reliable" datasets such as traffic counts, household surveys and more recently mobile phone data. These methods are typically expensive, time consuming and lack spatial/temporal coverage (Jestico et al. 2016). Crowdsourcing, a participatory method of data collection employing individuals/sensors, can be used to construct datasets. It is hypothesized that crowdsourced data have the potential to derive new insights within transport planning specifically perceptions of access and daily mobility patterns. However, there is a need to comprehensively examine the challenges and biases of using crowdsourced data to ensure it is fit for purpose. Given this, this research aims to answer the following research questions:- What understanding of mobility can be gained from crowdsourced data compared to conventional data sources?- Who contributes to crowdsourced datasets and how do their motivations affect data quality and spatial/ temporal biases?- What are the challenges and biases of crowdsourced geographic data in the context of mobility and how can the biases be mitigated and ensure data are fit for use?- How can crowdsourced data be reliably applied in transport modelling?The research questions will be achieved through examining the biases in crowdsourced data and their origins. An original framework will be devised to assess crowdsourced data quality and mitigate biases found. Finally, a case study will also be undertaken implementing crowdsourced data in an agent based transport model to determine novel guidance on how to suitably apply emerging crowdsourced datasets in transport modelling. This research provides a key chance to challenge what is considered 'everyday data' in transport modelling. Transport proposals have a key opportunity to be enriched through crowdsourced data, enabling a greater push towards 'data enabled decision making' an ESPRC research priority particularly in the Information and Communication Technologies theme. This research will consider the challenges of scalability, interoperability and reliability of crowdsourced sensor data, a key strategic focus in the EPSRC's ICT networks and distributed systems domain. It is stated that researchers should link research to real world tests, this research will apply the data in a transport model to highlight its fitness for use and real-world societal benefit.