Developing and exploring methods to understand human-nature interactions in urban areas using new forms of big data

这项研究的目的是探索如何利用手机空间大数据的新形式来研究城市中人与自然的互动。尽管人们越来越认识到绿色空间对健康和福祉的好处，但在过去一年半里，由于新冠肺炎疫情的限制，绿色空间的重要性变得更加突出。同样的这些限制也可能扩大了获得绿色空间方面的不平等，从而加剧了卫生方面的不平等。移动电话数据有可能更好地理解城市自然空间中的人类行为，但作为一种新的数据形式，它们也包含潜在的偏见。该项目探讨了我们如何克服这些偏见，并利用这些数据更好地了解城市地区的人与自然相互作用。这个特定的应用程序还可以被视为许多其他潜在应用程序的测试用例或演示，这些应用程序涉及对人口流动进行细粒度分析。背景：即使在“正常”时期，人与自然的动态关系对我们的城市也很重要。在大流行时期，自然空间的好处被放大，绿色空间在促进我们城市社会的健康和福祉方面发挥着更大的作用。在疫情期间，大自然一直是许多人身心喘息和营养的源泉，封锁规定增强了我们对当地公园和绿地的欣赏。这种与自然区域的更多接触很可能成为这个时代持久的遗产之一。然而，大流行病施加的限制（特别是对公共交通的限制）可能加剧了进入和使用绿色空间方面现有的不平等现象。传统上，抽样调查是了解绿色空间使用的最常用工具。它对于提供对自然空间的偏好和社会规范以及总体使用的变化的高层次图像仍然很重要。然而，样本量的限制意味着它不能详细了解不同类型的地点的使用变化，或者随着时间的推移，对相对短暂的行动限制的反应。不均匀的回复率或抽样策略的弱点也可能导致结果偏差。对于绿地管理者来说，调查无法提供特定地点的时空图景，无法为投资和管理战略提供信息，因为他们正在努力应对游客数量增加或疫情造成的其他使用变化带来的压力。由于可获得的数据量、广泛的人口覆盖范围以及提供的空间和时间细节，移动电话数据提供了巨大的潜力。然而，产生这些数据的过程往往相当不清楚，而且它们也可能包含人口覆盖方面的偏见，从而影响它们所提供的情况。我们需要密切关注数据的质量，并了解这种质量在不同的商业提供商之间是如何变化的。建议研究：我们将直接解决手机数据的偏见和代表性问题。我们使用的所有数据都是去识别的（即所有的姓名、电话号码或其他个人标识符都已被删除），但我们可以使用每部手机的移动来推断用户居住的地区，从而推断数据在地理和社会上的代表性。然后，我们可以调整或加权数据，以便在必要时提供更具代表性的图像。移动电话数据可以从不同的供应商那里获得许可，但对于商业运营商之间的数据差异几乎一无所知。我们通过比较两家不同的移动电话数据提供商的数据来探讨这一点。利用增强的数据集，我们将探索大流行不同阶段绿色空间使用模式的变化。我们还将研究使用不同类型网站的人的社会不平等，人们使用这些网站的频率和距离。

The aim of this research is to explore how new forms of spatial big data from mobile phones can be used to examine urban human-nature interactions. While the health and well-being benefits of greenspace have been increasingly recognised, they have taken on even greater significance over the last year and a half due to the Covid-19 restrictions. These same restrictions may also have widened inequalities in access to greenspace, and hence contributed to widening health inequalities. Mobile phone data have the potential to provide a better understanding of human behaviour in urban natural spaces but, as a novel form of data, they also contain potential biases. This project examines how we might overcome these biases and use these data to better understand human-nature interactions in urban areas. This particular application can also be seen as a test case or demonstrator for many other potential applications involving the fine-grained analysis of population mobility.Background:The human-nature dynamic is important for our cities even in 'normal' times. In a time of pandemic, the perceived benefits of natural spaces are amplified, with greenspace playing an even greater role in promoting the health and well-being of our urban societies. Nature has been a source of physical and mental respite and nourishment for many during the pandemic, with lockdown rules heightening our appreciation for local parks and greenspaces. This increased engagement with natural areas may well form one of the enduring legacies of this time. However, the restrictions imposed by the pandemic (notably on public transport) may have exacerbated existing inequalities on access to and use of greenspace. Traditionally, the sample survey is the most common tool for understanding the use of greenspace. It remains important for providing a high-level picture of changes in preferences and social norms towards nature spaces as well as overall usage. However, limitations of sample size mean it cannot provide detailed understanding of changes in the use of different kinds of sites or variations over time in response to relatively short-lived restrictions on movement. Uneven response rates or weaknesses in sampling strategies may also introduce biases in results. For greenspace managers, surveys cannot provide the kind of site-specific spatiotemporal picture needed to inform strategies for investment and management as they struggle to cope with the pressures of increased visitor numbers or other changes in use causes by the pandemic. Mobile phone data offer enormous potential by virtue of the volume of data available, the wide population coverage and the spatial and temporal detail provided. However, the processes by which these data are produced are often rather unclear and they may also contain biases in population coverage which impact on the picture they provide. We need to pay close attention to the quality of the data and understand how this quality may vary between the different commercial providers.Proposed research:We will address the issues of bias and representativeness in mobile phone data directly. All the data we use are deidentified (i.e. all names, phone numbers or other personal identifiers have been removed), but we can use the movements of each mobile phone to infer which area a user lives in and hence how geographically and socially representative the data are. We can then adjust or weight the data to try to provide a more representative picture if necessary. Mobile phone data can be licensed from different providers yet almost nothing is known about how data vary between commercial operators. We explore this by comparing data from two different providers of mobile phone data. With our enhanced datasets, we will explore variations in the patterns of greenspace usage across the different stages of the pandemic. We will also examine social inequalities in who uses different kinds of sites, how often and how far people travel to do so.