New perspectives on daily urban mobility: Harnessing the potential of Smart Card Travel data

英国和其他地方的许多城市都推出了智能旅游卡(STC)。这些数据构成了交通部门例行收集的一种新资源，补充了这些组织已经拥有的广泛的旅行数据。STC记录了公共交通中的所有电子客票交易，并以这种方式以前所未有的详细程度跟踪了大部分当地居民和游客的日常交通情况。与公共和私人领域持有的其他数据相结合，这些数据提供了以新的方式了解日常城市交通动态以造福于交通组织以及社会科学和政策中更广泛关注的重要问题的潜力。特别是，这些数据通过提供近乎实时的时间分辨率、地理细节和广泛的人口覆盖范围，补充了传统的、往往昂贵的数据源(英国人口普查中的上班旅行部分、家庭旅行调查中的旅行日记)。在收集TfWM提供的数据时(基于现有的正式数据共享协议)，这项研究将有助于在社会不平等和生活方式选择的背景下实现可持续和包容性流动的紧迫社会科学和政策问题(Kandt等人，2015年)。这项研究还设想提供一个框架，为社会科学研究更广泛地利用STC数据，从而为已经推出或将推出STC的其他城市制定蓝图。这项研究还直接促进了TfWM的目标，即利用其庞大的数据库支持其提供可持续和包容性移动性的任务。虽然STC数据特别令人感兴趣，因为它们提供了旅行需求的关键变量(例如，旅行生成、频率、距离、一天中的时间、居住时间)，但必须开发一些启发式方法，以得出有用的旅行指标，确定数据的覆盖范围，了解偏见的程度和操作，并将这些指标与更广泛的社会领域联系起来，如排斥或健康和福祉。出行需求是一种衍生需求，可以理解为各种日常活动的痕迹，这些活动在人口群体、社区或地区之间聚集在一起，很可能构成对城市动态的详细、多层次的了解。目前关于使用STC数据的工作仍处于初级阶段，特别是在将分析扩展到运输部门业务问题之外的努力方面。在这项研究中(并根据TfWM现有的利益)，需要确定在例行收集的数据集中代表了多少居民。这个问题与大数据中的偏见有关，因此与大数据分析中的一个常见问题有关。将数据转化为信息的具体方法包括以下策略和技术：(1)从旅行交易生成每日流动性概况，(2)将行程与日常活动联系起来，例如通过推断旅行目的地和目的，以及(3)从流动性概况推断人口统计数据。作为这些承诺的一部分，学生将需要通过STC记录的寄宿和GPS车辆跟踪数据的数据链接来利用数据中可用的隐含地理信息。因此，这项研究将长期确立的地理人口学与Paul Longley团队进行的地理空间研究联系起来(例如Longley等人2016年、Longley等人2015年)。这些步骤所产生的专业知识将使学生能够制定社会科学中的大数据分析研究框架，并应对交通部门的挑战。

Smart Travel Cards (STCs) are introduced in many cities in the UK and elsewhere. The data form a novel resource routinely collected by transport authorities, complementing the wide range of travel data already held by those organisations. STCs record all electronic ticket transactions in public transport and in this way trace daily mobility in unprecedented detail for a large proportion of both local residents and visitors. In combination with other data held in the public and private domains, the data offer potential to understand, in new ways, the dynamics of daily, urban mobility for the benefit of transport organisations as well as important issues of wider concern in social science and policy. In particular, the data complement traditional, often costly data sources (Travel-To-Work component of the UK Census, travel diaries in household travel surveys) by offering a nearly real-time temporal resolution, geographical detail and wide population coverage. In assembling the data provided by TfWM (based on an existing formal data sharing agreement), the research will contribute to pressing social science and policy questions of achieving sustainable and inclusive mobility in the context of social inequalities and lifestyle choices (Kandt et al 2015). The research is also envisioned to deliver a framework to harness STC data for social science research more generally and thereby develop a blueprint for other cities that have introduced or will introduce STCs. The research also contributes directly to the objectives of TfWM of using their vast databases to support their mandate of delivering sustainable and inclusive mobility. While the STC data are particularly interesting because they deliver key variables of travel demand (e.g. trip generation, frequency, distance, time of day, dwelling time), a number of heuristics will have to be developed in order to derive useful indicators of travel, ascertain the coverage of the data, understand the extent and operation of bias and link the indicators to wider social domains, such as exclusion or health and well-being. Travel demand is a derived demand and may be understood as traces of all sorts of daily activities, which aggregated across population groups, neighbourhoods or regions may well add up to a detailed, multi-level understanding of urban dynamics. Current work on using STC data is still in its infancy, especially in relation to efforts to extend analytics beyond operational questions of the transport sector. Within the research (and in accordance to existing interests at TfWM), it will need to be established, how many residents are represented in routinely collected datasets. This question is related to bias in Big Data and hence relates to a common concern in Big Data Analytics. Specific methods of turning the data into information include strategies and techniques to (1) generate daily mobility profiles from travel transactions, (2) link trips to daily activities, e.g. by inferring trip destinations and purposes and (3) infer demographics from mobility profiles. As part of these undertakings, the student will need to make use of the implicit geographic information available in the data through data linkage of STC-recorded boardings and GPS vehicle tracking data. The research thus links to the long-established geodemographics with geo-spatial research undertaken by Paul Longley's team (e.g. Longley et al 2016, Longley et al 2015). The resulting expertise from these steps will enable the student to both develop research frameworks of Big Data Analytics in social science and address challenges in the transport sector.