By Peter Rauxloh, Director of Technology Services at Museum of London Archaeology (MOLA).
This article is part of our series: a day in the software life, in which we ask researchers from all disciplines to discuss the tools that make their research possible.
In 1746, John Roque, a French Huguenot émigré and enthusiastic surveyor for hire, published his iconic map of London. It is a valuable representation of London at the time, and is now the focus of a new web resource. Locating London’s Past allows you to search a wide body of digital resources relating to early modern and eighteenth-century London and to map the results.
The map was first published in the form of 24 sheets at a scale of 26 inches to the mile. It is an important document in itself, but even more so when its position in the development of modern surveying technique is considered. Rocque was working at a time when triangulation was being adopted by surveyors (triangulation is the practice of making multiple observations of horizontal angles between survey points to build a network of points with an identifiable level of accuracy). Rocque adopted these methods while he carried out his survey after he realised that initial measurements carried out at street level were not accurate enough.
Locating London's Past uses an intuitive geographic information system (GIS) interface to map and visualise textual and artefactual data as it relates to 17th and 18th century London. The Museum of London Archaeology Service's (MOLA) created a seamless mosaic of the entire area from scans of the original map. This required the removal of sheet edges and margins and the application of, often minute, scale factors to allow for 250 years’ worth of paper shrinkage. The original map was made up of individual sheets, and occasionally one can see the different approach of different engravers in two halves of a field separated by a sheet boundary: with one preferring stipples and the other championing the wavy line. Finally, a geographically accurate place name index was created for London in 1746 which associates the orginial map with an existing digital resource that recorded the location of crimes which were tried at the old bailey between 1674 and 1913.
GIS-based methods were developed to yield a geographically accurate point data set for each combination of street or place and the parish and ward in which they existed. At a time when street numbers were not yet common, this combination was generally used to reference a place in contemporary documentation. This stage used the Rocque map as a guide to build a geographically accurate street network as it existed at the time. GIS tools to auto-derive an accurate polygon for each street and place, populate it with a name, intersect those named polygons with parish and ward polygons, and derive from these the final product: a point centroid for each resultant polygon.
Figure shows an area west of the City with the colour coded lines being streets digitised from the later, more accurate mapping. This is overlain over the original Rocque map – hence the appearance of discrepancies. Orange represents other identified areas.
The creation of this street network used Rocque’s map as a guide as to which streets and places actually existed in the middle of the 18th century, located those streets on the more accurate geo-referenced ordnance survey 1st and 2nd edition maps of some 150 years later, and then digitised those features from the later resource into a GIS vector data set. Since one ought to be able to get from any one street to any other – even in the 18th century – this data set will form a connected network, which provides a useful check on the work. For the streets, each was allocated a rough width and a position in a 7 tier hierarchy, and then all street centre lines were converted to polygons using a buffer based on their width.
Next, a clipping routine was repeatedly applied so that the overlaps between streets were removed with more major streets clipping the end of lesser thoroughfares which joined them. A simple polygon centroid tool then created a point for each street. The final act was to take polygons representing the parish and ward boundaries extant at the period, and use these to spatially intersect the clipped street polygons, producing a new set of areas with the attribute of street name and parish/ward in which they exist. From these new polygons, final centroids were created: one point for every combination street/place and parish/ward.
The project provided historical insights and methodological benefits, an observation made possible when the mosaic of Rocque’s map was geo-referenced for the entire area, showing how his faith in trigonometry had been borne out. A high degree of articulation with modern mapping observable in the historic core, got progressively worse as one went out from there. The tall buildings of the City provided a target rich environment for the keen surveyor to check and re-check their angles, while the less developed, lower height of buildings in the extramural area, provided less scope for such checks and thus the detection and removal of error. The production of a geographically accurate historical street network has formed the basis for producing similar products from earlier and later mapping, where the addition or removal of new features is required rather than starting from scratch. The exercise also yielded a form of local georeferencing to enable features to be digitised directly from the older maps in cases where they had been obliterated by the time the newly formed Ordnance Survey, rolled into town.
A fuller account of the work is available on the Locating London Website.