Can anyone help me link the approximately 15,000 Parish-Level records from aggregating census variables from the full-count Great Britain 1851 census on the Parishes with the relevant parish polygons in the Parish-level shapefile I have as a registered user on the UK Data Service platform, but note this available shapefile has no ID values for each parish polygon that match the ID values for parishes in the IPUMS NAPP full count Great Britain 1851 census. But, the Names of each parish are available in both datasets, but there are many duplicates or parishes in different parts of England with the same name. However, within each District there are likely not so many duplicates to be an issue. Is there way to match aggregations to the parish level from this 1851 census for various variables to the most likely correct parish polygon in this shapefile by first matching District names and then the Parish names?
Thank you, Karl
Unfortunately IPUMS does not have any shapefiles or other information on the geographic boundaries of the parishes that are listed in the 1851 U.K. census. The geographic variables available for the 1851 UK full count data are:
My understanding of your issue is that you want to link information about parishes from IPUMS to the geographic boundaries of the parishes as given by the UK Data Service. The parish codes in the IPUMS data differ from those in the parish shapefile. It’s not simple to match them on name because there are multiple duplicate names. You have district in the shapefile so you can narrow down some of the parish matches within a district, but don’t know how to match the names.
First, I would recommend reaching out to the UK Data Service and/or the researchers who created the shapefile you are using. They may be able to tell you how they coded the parishes and help you determine how to make these matches.
Assistance with matching or giving analytical advice is beyond the scope of IPUMS User Support. You could start by reading about fuzzy matching in R here. Much historical data work involves some type of matching and you should be able to find many resources and papers about matching techniques online.
Hi Isabel,
Thanks for these ideas, I think if I had time I might just locate the centroid of each Parish in the 1851 British Census, and then just make a Thiessen Polygon coverage for all of England and Wales with the roughly 15,000 points, and clip to the coastlines and Scotland boundary. In the Parish shapefile available on the UK Data Service, there are about 23,000 Parish polygons.
I would think within each Sub-District there would not be many Parishes with the same name, and in those cases it would probably be possible to figure out which is which based on comparing the total number of people in 1851, or some key variable like the number of miners or farmers.
If such a historical GIS dataset was made, could I make it available Open Access either through IPUMS, or from a Dataverse? In this case the Thiessen Polygon shapefile I made would be solely my creation, as I would have obtained the centroids of each Parish poly from Google Earth and not the UK Data Service shapefile. While the Parish level 1851 census data aggregations would be from IPUMS North Atlantic Project which appears to also be Open Access.
Best, Karl
Estimates derived from IPUMS International may be redistributed according to our Terms of Use. However, the underlying microdata from IPUMS International cannot be redistributed for any reason.
We would be excited to see the outcome of this work. If you share a link once it is finished, we would gladly highlight the availability of this resource to others who would benefit from the work.