knitr::opts_chunk$set(echo = TRUE)

Three Classes of Landmarks

1) Squares (road entrances) - Golden Square and Soho Square.

2) Label at location - Karl Marx, John Snow, The Pantheon.

3) Label separate from location - St James Workhouse, St Lukes (Berwick)

https://www.british-history.ac.uk/survey-london/vols31-2/pt2/pp219-229#h3-0012 https://www.londonpicturearchive.org.uk/view-item?i=19777&WINID=1698775877925

Methodology

The geographic or "latlong" coordinates for 'landmarks' dataset were computed by georeferencing the nominal coordinates using QGIS. Partly as a check (validation) and partly as an experiment, the geographic coordinates for 'landmarksB' were computed by using ratios and proportions. I did this because georeferencing seems add a bit of "noise" to the relative position of points (i.e., nominal v. geographic). A good example is that the stacks of cases ("dots" or "bars") typically lose their alignment in the process.

In Dodson and Tobler's digitization, roads and street are represented by a set of straight line segments. The ratio and proportion method works by defining a landmark's address not as an x-y coordinate but in terms of its relative position along the line segment where the landmark is located.

For example, Agyll House is located on Argyll Street, which is composed of three line segments. The building's frontage appears to be on the Southern most segment ("162-1"). A reasonable and convenient choice for its address is the mid-point of the segment. Since I have already georeferenced all the road segments, my thinking is that I can use this ratio to "compute" the geographic coordinate of Argyll House without having to georeference it. I just compute the mid-point of geographic coordinates ("latlong") of segment "162-1". In this example, "162-1" is approximated 80 meters in length. Thus the geographic coordinate is 40 meters from either end points along the segment.

This saves an additional georeferencing step and is potentially more "accurate". This is an experimental, work in progress.



lindbrook/cholera documentation built on Jan. 13, 2025, 3:49 p.m.