Assessment of Landownership Change using Registers of Scotland Data
What were we trying to find out?
We were trying to determine how useable Registers of Scotland (RoS) data was for research and policy implementation purposes. Landownership data in Scotland is limited, and therefore being able to use the official land register would unlock new means of exploring issues relating to landownership. More specifically, we wanted to demonstrate if two of RoS’s commercially available data sets (Non-Residential Land Values and INSPIRE) could be joined to show the spatial extent of sales and how accurately they ‘match’ to each other.
What did we do?
We acquired data from 2018 to 2023 and mapped these across three case study areas. An automatic ‘matching’ process was created and used to match the two datasets together (enabling the spatial extent of each sale to be seen), with the accuracy of the match being evaluated. Manual checking was also required to fully confirm the matching accuracy, and the methodology for this process is included in this report.
What did we learn?
We learnt that using RoS data in this way is problematic and attempts to automate the process were not fully successful. However, after manual checks were conducted, the data did provide some use cases and we were able to conduct some analysis on landownership and land transactions.
Concerning the validation of matching accuracy, of the 4875 Non-Residential Land Value applications only 57.83% were correctly matched automatically. An additional 10.28% of the total number of applications were also made available for analysis once they had been resolved through manual corrections. The main challenges to the matching process included missing coordinates and ownership area within the Non-Residential Land Values dataset and unavailable polygons within the INSPIRE dataset, although other unexpected issues were also encountered.
Regarding the three case study areas. Spey Catchment: 16,561.81 ha transacted, dominated by infrequent large estate sales. Shetland Islands: 14,391.15 ha transacted, with a high frequency of smallholder sales. GASABR: 81,239.93 ha transacted, mostly involving mid-sized parcels.
What do we recommend and what happens next?
We provide two potential avenues for policy. For our project, the next stage is to overlay our ownership and transaction GIS layers with other spatial datasets concerned with financial flows (grant payments, agricultural subsidies etc.) into land.
