Methods for analysis of space-time data: a Bayesian approach

时空数据分析方法：贝叶斯方法

• 批准号: RGPIN-2014-06359
• 负责人: Law, Jane
• 金额: $ 1.46万
• 依托单位: University of Waterloo
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2015
• 资助国家: 加拿大
• 起止时间: 2015-01-01 至 2016-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=588960
• 关键词: Methods; analysis; space; time; data

The proposed research program will address current concerns and raise the quality of geographic research in Canada. Popularity of web technology and mobile devices, such as iPhone or iPad, has posed new challenges to many disciplines in Canada and elsewhere. Data associated with individuals or events can be collected anytime. They contain information about where (geographic coordinates or addresses) and when (hour, minute or second) exactly an event (what) occurred. We are aware that the information can be used to generate knowledge about events that are or will be happening around us at any point in time, which could have a ground-breaking impact on our life, yet we do not know how. Methods developed for analyzing data that contain precise location and time information, referred to as space-time (ST) data, have focused on environmental and climatic sciences, but less so on social, health, or economic sciences. This proposed research program aims to address the above concerns. Bayesian, a relatively new approach of geographical analysis will be adopted to develop statistical methods for analysis of ST data in social, health, or economic sciences. The proposed program consists of short and long term goals of research and training. Short-term goals are to develop quantitative methodologies that would enhance current capabilities to analyze ST data. For instance, in terms of safety, the methods developed will enable the following questions to be answered accurately: Are changes or trends of violence crime in my local areas significantly different from others? If so, what are the associated (local) risk factors that might have caused the difference? Where are the areas that are significantly less safe at a specific season, month, day of the week or hour of the day? What is the probability of my neighbourhood being burgled in the summer? Currently, it is not possible to provide accurate and statistically-sound answers to the above questions. Long-term goals are to extend the ST methods to enable real-time analysis. Questions that can be addressed then will include what is the chance of being a victim or to be rescued at my current location and time (in case of a disaster) according to the real-time response from my smartphone? Through my current NSERC grant, I have established a research team to develop methods for spatial and ST analysis. Our team has published over ten peer-reviewed journal articles on related methods in the past three years. Upon renewal of the grant, more students will be trained to develop and apply the methods using Canadian data. A research centre on ST analysis in Canada could materialize in the longer term. Novel methodologies for analysis of ST data resulted from the research could transform the approach that researchers in many disciplines use and analyze geographic information. Many fields are facing the “Big Data Challenge” – the need for new methods of data analysis including ST analyses is compelling. Some practical significance of the program include 1) enable statistically-sound analysis of ST data; 2) prepare for the changing methods of data collection by mobile devices including the “Google glass”; and 3) Canada is no longer relying on the census long form to collect socio-economic data. Population or other surveys can be conducted via mobile devices to collect ST data that can be analyzed using the methods developed. Surveillance systems for public health and safety in situations of flooding, earthquake, or other crisis can benefit from such modern ways of data collection provided that sound methods of ST analysis are in place. Apart from geographic information science and other fields in Canada that conduct geographical research, global health and safety will advance through the research.

拟议的研究计划将解决当前的关注点，并提高加拿大地理研究的质量。 Web技术和移动设备（例如iPhone或iPad）的受欢迎程度为加拿大和其他地方的许多学科带来了新的挑战。可以随时收集与个人或事件相关的数据。它们包含有关地理位置（地理坐标或地址）以及（小时，分钟或第二个）完全发生事件（什么）的信息。我们知道，这些信息可用于在任何时间点在我们周围发生的事件产生知识，这可能会对我们的生活产生突破性的影响，但我们不知道如何。为分析的数据开发的方法，其中包含精确的位置和时间信息（称为时空数据（ST）数据），重点关注环境和文化科学，而不是社会，健康或经济科学。该建议的研究计划旨在解决上述问题。贝叶斯（Bayesian）将采用一种相对的地理分析方法来开发统计方法，以分析社会，健康或经济科学中的ST数据。 拟议的计划包括研究和培训的短期和长期目标。短期目标是开发定量方法，以增强当前分析ST数据的能力。例如，就安全性而言，开发的方法将使以下问题准确地回答：我当地的暴力犯罪的变化或趋势是否与其他问题有很大不同？如果是这样，可能导致差异的相关（本地）风险因素是什么？在特定的季节，一个月，一周或一天中的一个月，一天或小时的特定季节，在哪里安全的区域？夏天，我的邻居被盗窃的可能性是多少？目前，无法为上述问题提供准确且统计上的答案。长期目标是扩展ST方法以实现实时分析。然后可以解决的问题将包括成为受害者的机会，或者根据我的智能手机的实时响应，在我当前的地点和时间（如果发生灾难的情况下）被救出的机会是什么？ 通过当前的NSERC赠款，我建立了一个研究团队来开发空间和ST分析的方法。在过去三年中，我们的团队发表了有关相关方法的十篇经过同行评审的期刊文章。续签赠款后，将培训更多的学生使用加拿大数据来开发和应用这些方法。从长远来看，加拿大的ST分析研究中心可以实现。 研究产生的ST数据分析的新方法可以改变许多学科中研究人员使用和分析地理信息的方法。许多领域都面临着“大数据挑战” - 对包括ST分析在内的新方法分析的需求令人信服。该计划的某些实际意义包括1）对ST数据进行统计分析； 2）为包括“ Google Glass”在内的移动设备收集数据收集的不断变化的方法； 3）加拿大不再依靠长期的人口普查来收集社会经济数据。可以通过移动设备进行人口或其他调查，以收集可以使用开发的方法来分析的ST数据。在洪水，地震或其他危机的情况下，针对公共卫生和安全的监视系统可以从这种现代数据收集方式中受益，前提是有ST分析的合理方法。除了进行地理研究的地理信息科学和加拿大其他领域外，全球健康与安全还将通过这项研究前进。