Heritage Data Analytics: Sustainable strategies for large and complex stratigraphic and chronometric data.

点击翻译按钮获取中文摘要

Our chronological understanding of archaeological sites develops gradually during field and post excavation work. First, we understand the relative chronological and contextual information, as represented by the stratigraphy and cultural finds respectively. Later, scientific (eg radiocarbon) dates are obtained and, finally, statistical modelling is used to draw everything together. At present this final stage is extremely laborious and requires considerable expertise because no tools exist to automate anything except the final statistical calculations. The proposed PhD student will develop numerical, analytical and graphical tools to improve this situation, with the goal of semi-automating chronology construction.The student will begin by characterizing the most common types of data that heritage practitioners and researchers create and work with, and particularly those created by fieldwork investigations. They will consider how the re-use of such datasets could be made easier and more effective by enhancing or developing better standards for digital data archives of excavation records. Working with Historic England, the student will then focus on protocols and algorithms for interfacing with current and legacy site databases, while Buck and others will work on improvements to the modelling software. Working in this extended team of archaeologists, modellers and programmers, the student will learn techniques and protocols from field and laboratory archaeology, software engineering, graphical and statistical modelling and Bayesian inference.The project is important because it addresses growing problems caused by a lack of standardized approaches to the archiving of excavation data, especially key stratigraphic and phasing data, often held in hard-copy matrix diagrams or unstructured database tables. The PhD will help inform decisions on digital archiving standards and best practice for stratigraphic data deposition and re-use, while developing our understanding of how statistical and technical approaches can best address the problems involved. It will provide a better understanding of the user requirements for interface tools to make analysis of large sites (especially legacy sites) practical.Dye and Buck (2015) have recently shown that at least semi-automation of the link between archaeological data and Bayesian modelling is practical. This PhD will extend the tools they have begun to develop, to accommodate a wider range of archaeological site database protocols and complement the the developments in Bayesian chronological modelling being undertaken by other members of Buck's team.Key research questions to be addressed include: What are the most common types of data that heritage practitioners create and work with? How can those various types of data be best characterised? How might adoption of better standards for digital archives of excavation records make re-use of such data sets easier and more useful? How might approaches to excavation data recording and archiving be improved to better enable the use of Big Data technologies?The proposed PhD research is highly interdisciplinary, and so the student will spend the first year reviewing literature and data (with Historic England support) while learning the basics of several specialist computer protocols and languages (with the support of Buck). In years 2 and 3, the student will work closely with Buck, Ayala and Dye, extending or rewriting the inputs to the prototype software written by Dye (to accompany Dye and Buck, 2015), interfacing it to key file formats and protocols identified in year one. The software developed by Dye is only a prototype and far from fully functional, so thestudent will have considerable scope to decide exactly how and in what way the PhD should progress. Dye T.S. & Buck C.E. (2015). Archaeological sequence diagrams and Bayesian chronological models. Journal of Archaeological Science , 83, 84-93.