3. Data Sources & Methodology
The Scottish Government Urban/Rural Classification 2013-2014 was created by combining population and accessibility information to distinguish between urban and rural areas across Scotland. Population information is sourced from the Settlements dataset provided by National Records of Scotland (NRS), and accessibility information is obtained by calculating drive times from the centres of Settlements with a population of 10,000 or more (i.e. urban areas). Table 3.1 summarises the datasets used to create the classification.
Table 3.1: Data Sources for the Scottish Government Urban/Rural Classification
| Dataset || Source |
| Scottish Settlement boundaries, centroids and population estimates || National Records of Scotland (NRS), 2012 version (released July 2014) |
| English Settlement centroids for populations of 10,000 or more (i.e. Berwick-upon-Tweed and Carlisle) || Ordnance Survey (OS) 50K Gazetteer, 2014 |
| Integrated Transport Network (ITN) for Scotland, major routes for Northern England || Ordnance Survey (OS) MasterMap (ITN) 2014 |
| Scottish Ferry Routes || Scottish Government, 2014 |
| High and Low Water Mark coastline boundary || Ordnance Survey (OS) BoundaryLine, 2014 |
Settlements define the built-up areas in Scotland that are generally more identifiable as the traditional towns and cities than administrative boundaries such as Council areas, much of which consists of land that is not developed and unpopulated. The Settlements dataset is produced by NRS every two years, with the current version (Settlements 2012) having been released in July 2014. Small Area Population Estimates (SAPE) together with information from the Royal Mail Postal Address File (PAF) were used to classify 2012 postcodes into high or low density, and this information is then used to identify contiguous postcodes with a total population of 500 or more that make up a Settlement. For more information on how Settlements are defined, see the NRS website at http://www.nrscotland.gov.uk/.
Settlement centroids for Carlisle and Berwick-upon-Tweed in Northern England were also included in the analysis as they have populations of over 10,000 and are within a 30 minute drive of the Scottish border, and thus may influence accessibility results. These centroids were obtained from the Ordnance Survey's 50K Gazetteer and adjusted slightly such that they fall along the road network.
Road and Ferry Network
For the calculation of drive times, a network dataset of the transport network in Scotland needed to be created. Required inputs for the travel network were the Integrated Transport Network (ITN) for Scotland with a 10km buffer to include ITN for Northern England, and vehicular ferry routes in Scotland. ITN is the definitive, most accurate and up to date geographic reference for Great Britain's road structure, and is provided by the Ordnance Survey (OS) as part of their MasterMap product.
A further input to the network dataset was Scottish Vehicular Ferry Routes. The Ferry Route dataset was created by the Scottish Government by surveying online timetables and maps from the individual service providers (both private and subsidised). Routes were digitised against Ordnance Survey background mapping. A nominal ferry speed of 15 km per hour was applied. An additional 30 minutes was also added to the travel time figures to account for wait time prior to boarding.
Lastly, the High and Low Water Mark Coastline boundary originates from OS BoundaryLine data and was used to clip the road grid and final datasets.
The first stage in creating the classification was to categorise the Settlements dataset using the population thresholds of 125,000, 10,000 and 3,000 to identify those settlements from which drive times will be calculated. Settlements were grouped into the following categories:
(1) Large Urban Areas - populations of 125,000 or more
(2) Other Urban Areas - populations of 10,000 to 124,999
(3) Small Towns - populations of 3,000 to 9,999
(4) Rural Areas - populations less than 3,000
The next step was to distinguish between accessible and remote areas. This was done by calculating a 30 minute drive time from the population weighted centroid of Settlements with a population of 10,000 or more (i.e. Large and Other Urban Areas, and including the two settlements in northern England of Berwick-upon-Tweed and Carlisle). For the 8-fold Urban/Rural Classification, an additional 60 minute drive time was also calculated. Thus, the following definitions of remoteness were defined:
(1) Accessible - areas within a 30 minute drive time of a Settlement with a population of 10,000 or more.
(2) Remote - areas that are more than a 30 minute drive time (6-fold classification), or areas that have a drive time more than 30 minutes but less than or equal to 60 minutes (8-fold classification) from a Settlement with a population of 10,000 or more.
(3) Very Remote - areas that are more than a 60 minute drive time from a Settlement with a population of 10,000 or more (8-fold classification only).
The drive time analysis was performed using the specialist GIS (Geographic Information System) software, ESRI ArcGIS Network Analyst Extension. Firstly, a network dataset was built from the road and ferry networks, and each road type was classified by an average speed, shown in Table 3.1. Settlement boundaries identified whether the rural or urban speed was applied. For example, motorways would have been assigned an average speed of 104.6 kph in rural areas, and 70.8 kph in urban (built up) areas. The speeds for each road type are the average for that class but it should be noted that the process does not take into account peak and non-peak travel times.
Table 3.2: Road classes and average speed applied in the classification
| Route Type || Rural Speed (kph) || Urban Speed |
| Motorway || 104.6 || 70.8 |
| A Road || 64.4 || 30.6 |
| B Road || 54.7 || 25.8 |
| Minor and Local Road || 40.2 || 22.5 |
| Alley || 40.2 || 22.5 |
| Pedestrian Road || 0 || 0 |
| Private Road - Publicly accessible || 40.2 || 22.5 |
| Private Road || 40.2 || 22.5 |
| Ferry || 15 || 15 |
Once the road/ferry network dataset had been created and reclassified in terms of average speeds, the drive time analysis could be calculated from those settlements with a population of 10,000 or more. Outputs of the analysis were boundary extents of both 30 and 60 minutes drive time. Each settlement was initially classed depending on its estimated population, but with the creation of the drive time extent layers, they were further classified in terms of accessibility. Accessibility categories were assigned to the settlement boundary layer based upon the location of the settlements' population weighted centroid. Classifying the settlements by their centroids means that the entire settlement will be assigned to a single class, regardless of whether the area is split by the drive time extent boundary.
Finally, the re-classified settlement boundaries and drive time datasets were combined to form one national dataset containing the 2, 3, 6 and 8-fold urban/rural definitions. All layers were clipped to both the high and low water mark coastline.