This paper describes the ongoing creation from historical primary sources of “EMEW”, an online Gazetteer of Early Modern England & Wales. The speaker is Technical Director of the World Historical Gazetteer, and was (as an early modern historian) co-leader of a public history project, “Viae Regiae”, which crowd-sourced the transcription and geolocation of various early modern maps and texts. EMEW is intended as both a qualitative and a quantitative resource for economic and transport historians interested in the spatial development and decline of commercial networks over time. The paper gives a detailed account of the diverse computational processes and tools employed in the preparation of the various Linked Open (Geo)Datasets which will constitute EMEW.

At the core of EMEW will be a digital rendition of “Index Villaris”, a list of some 24,000 place names first published by John Adams in 1680, and the first gazetteer of England and Wales to attempt the inclusion of geo-coordinates. The AI-based transcription tool “Transkribus” was first trained to recognise the idiosyncrasies and symbology of the published source. The resulting transcript was then corrected manually, and GIS techniques were used to transform Adams’ coordinate reference system (CRS), aligning the place names more closely with their known modern locations. The entire dataset was then processed using “Locolligo”, a tool developed by the speaker for a project based at the British Library, which facilitates the geospatial linking of datasets. Index Villaris place names were thereby linked as far as possible to Wikidata, to Open Street Map (OSM) road nodes, and to place names from the GB1900 Gazetteer. The GB1900 Gazetteer is itself a crowd-sourced dataset comprising all of the textual content of the Ordnance Survey six-inch to the mile maps of England, Scotland and Wales, dating from 1888-1913, and amounting to some 2.5 million names.

Supplementing Index Villaris, EMEW will include datasets derived by Viae Regiae volunteers from Christopher Saxton’s county maps (published in the 1570s) and from records of John Leland’s travels in the 1530s. These were prepared online collaboratively using “Recogito”, a geolocation and transcription tool that employs IIIF for serving both maps and texts.

EMEW is to be bolstered further by the inclusion of a dataset which geolocates all of the monasteries closed during the English Reformation, adding temporo-spatial dimensions to any study of this period of especially-significant economic upheaval. A dataset of markets and fairs in England and Wales, originally covering the period to 1516, is being extended to 1900 through traditional archival research, exploring Crown patents and court records held at the UK National Archives. Yet another dataset originating in a War Office survey of accommodation and stabling made in 1686 will add a qualitative proxy-dimension to the market towns recorded in Index Villaris six years earlier. Together, these datasets will help with charting the changing economic status of both localities and regions.

The paper will also describe the speaker’s work on “desCartes”, a computational workflow which employs AI, computer vision technologies, and Python notebooks to attempt the extraction of road vectors from historical maps. Using a combination of random forest and convolutional neural networks for pixel classification, the initial aim is to produce a vector road map from the same set of Ordnance Survey maps as was used to produce the textual data of GB1900, which were surveyed before the commencement of major road-building programmes. Through further research identifying roads built following the legally-authorised compulsory purchase of land or those created by the documented process of enclosure, this vector road map will form the basis for historical regression. Ultimately, it is hoped that a digital map will be produced representing speculatively the development of the road network of England and Wales since 1540, as both a driver and indicator of historical economic development.

World Historical Gazetteer (WHG) provides a collection of content and services that permit world historians, their students, and the general public to perform spatial and temporal reasoning and visualisation in a data rich environment, at local, national, global and trans-regional scales. Among the array of WHG services is the “Dataset Collection”, which allows users to group together datasets which share a common theme. This paper concludes with an account of how this feature is being used to create EMEW.

Many of the key themes of the 5th Spatial Humanities Conference are addressed by this paper. In particular, it highlights the use of gazetteers, artificial intelligence (computer vision and deep learning), spatial explorations of narratives, GIS and spatial statistical analysis, spatial connections and networks, linking maps and texts, geospatial data enrichment and annotation, historical maps and georeferencing, Linked Open (Geo)Data, IIIF applications, labs notebooks, workflows and infrastructure, data mining, visualisation, and the challenges of geolocation.