The production of geographic data has traditionally been the purview of institutions such as the Ordnance Survey and US Geological Survey. The past three decades have seen a technological revolution brought about by mobile computing resources and the World Wide Web. Ordinary citizens now have the tools to produce geographic information en-masse. OpenStreetMap, one of the world’s most important and extensive geographic datasets has arisen out of this Volunteered Geographic Information phenomenon. Freely available to be both used and produced by ordinary people, it represents a paradigm shift which has changed our relationship with geographic information.The free and open nature of OpenStreetMap has given rise to novel and often mission-critical uses, often among people with little or no interest in traditional quality assurance frameworks. The shift away from authoritative data sources and traditional quality assurance paradigms raises problems for geospatial data consumers who still need to make informed trust judgements. In a milieu characterised by a diverse and dynamic range of use cases, large volumes of data and no established quality assurance paradigms, we need new ways of understanding Volunteered Geographic Information. One of the most difficult components to document all the contributors and their contribution practices, which operate at a scale and diversity not found in traditional science.Provenance data encodes much of this information and is useful for providing localised data documentation. Provenance Network Analytics is a methodology which has the potential to provide a principled automated means of analysing provenance data at scale. However, it has only been implemented in relatively simple, smaller scale use cases. OpenStreetMap does not explicitly record provenance data. There is also no framework for understanding of the necessary measurement and interpretation strategiesIn this thesis we address these issues by providing a novel method of provenance reconstruction which produces a provenance dataset in an interoperable standard. We provide a framework for provenance measurement using metrics which allow the analysis of large volumes of data. Using OpenStreetMap provenance extracted from the Southampton area in the UK, we conduct a descriptive analysis of OpenStreetMap provenance data. The results provide an understanding of the drivers of variation in OpenStreetMap contribution practices. This work repositions VGI provenance as a new and novel form of geographic data which can provide insights into the nature of Volunteered Geographic Information and the human and physical environment it describes.