Scottish Vacant and Derelict Land Survey: quality assurance process

This note sets out the quality assurance checks that are undertaken to produce the Scottish Vacant and Derelict Land Survey.

Data checks

The survey is returned in the form of two Excel spreadsheets which detail the vacant and derelict sites for each Planning Authority. Excel file SVDLS-A details new sites and SVDLS-B reused sites. The variables referenced below are explained in more detail in Annex A. Planning Authorities also provide shape files detailing new sites in a geospatial format that allows them to be mapped.

Excel checks

These check are undertaken using a macro and are made on the SVDLS-A dataset (new sites) and the SVDLS-B dataset (reused sites). Many of the check below correspond to variables on the Excel survey form as detailed in Annex A.

SVLDS-A dataset

The following checks are made:

• check data present
• COUNCODE Valid
• SITECODE present
• DBHist valid
• name present
• address present
• site size >= 0.1 and < 1000
• site type valid
• owner valid
• year surveyed valid
• TimeVD valid
• PrevUse valid
• derchar valid
• intrlocn2 valid
• devpot valid
• OSGRID check or eastings/northings present.
• east coord right length
• north coord right length

SVDLS-B error checks

The following checks are made:

• check data present
• COUNCODE Valid
• SITECODE present
• site size >= 0 and < 1000
• NewUse valid
• mixed use valid
• NLUC valid
• inuse dates valid
• funding valid
• VDLF valid
• OSGRID check or eastings/northings present.
• east coord right length
• north coord right length
• invalid length for OSGRID
• duplicate site codes for the authority – These might be errors or could be genuine duplicates where a large site has been split into several pieces, with 1 plot of land going to new housing, another going to offices etc. The respondent may choose to put several records in the B dataset with the same sitecode showing the different new uses for the land. Review and check with LA if necessary

SAS checks

These checks are run using SAS programming and are made on the SVDLS-A dataset (new sites) and the SVDLS-B datasets (reused sites). The check correspond to variables on the Excel survey form but unlike the Excel checks are made using comparison with historic data as well.

SVLDS-A dataset

The following checks are made:

• invalid council codes (COUNCODE)
• missing site codes (SITECODE)
• invalid code for a site’s database history (DBHIST can only be 1 or 2)
• missing name and address
• questionable site size (0.0999 < SITESIZE < 999.99)
• invalid site type (SITETYPE)
• change in site type since previous year
• invalid ownership (OWNER1/ OWNER2)
• invalid date of first inspection? (1979 < INSPYY<= current year)
• invalid code for the time the site has been vacant/derelict? (TIMEVD)
• new site became vacant or derelict prior to previous year – check with LA if they want this included in historical data
• invalid code for previous use of the site? (PREVUSE)
• invalid code for the derelict characteristics of a site (DERCHAR)
• where site has a derelict characteristic, is it site type derelict (i.e. SITETYPE needs to be equal to 11)
• invalid code for the settlement-based location of the site? (INTRLOCN2)
• location has changes since previous year
• vacant sites (SITETYPE = 21,22) classed as countryside (INTRLOCN2 = 3). Vacant sites are only valid for the survey if they are in settlements
• invalid code for the development potential of the site? (DEVPOT)
• missing East, North coordinates and no alternative OS grid ref
• invalid length for East, North coordinates (this may flag up North co-ordinates in Orkney. Shetland that can have an extra digit (7 rather than 6), code still needs to be revised to cope with this
• invalid length for OSGRID
• duplicate site codes for the authority
• sites where co-ordinates have changed since last year – these are not necessarily errors as minor corrections may have been made or the site might have changed size so the centroid has moved. This list is just to help identify sites to review manually when doing the GIS check for those authorities with lots of sites
• check new sites weren’t on the SVDLS register last year

SVLDS-B dataset

The following checks are made:

• invalid council codes (COUNCODE)
• missing site codes (SITECODE)
• invalid site size (SITESIZE > 999.99 or < 0).
• invalid code for the new use of the site? (NEWUSE).
• sites removed for definitional reasons? ( NEWUSE = 34).
• invalid code for whether or not the new use of the site is mixed?
• invalid code for National Land Use Classification (NLUC)?
• invalid month or year brought back into use
• brought back into use before the previous year – need to check with LA to see if the historical files should also be updated.
• invalid code for the funding mechanism involved in the site’s new use
• invalid code for VDLFund
• missing East, North coordinates and no alternative OS grid ref.
• invalid length for East, North coordinates (this may flag up North co-ordinates in Orkney. Shetland that can have an extra digit (7 rather than 6), code still needs to be revised to cope with this
• check re-used sites were on the SVDLS register in previous year

SVLDS-A and B dataset cross checks

The following checks are made:

• check changes in site area. Where a site is new last year, it can be split this year creating a new smaller site and a reused site. Both site sizes combined should equal the size of the whole site last year
• compares overall site additions and removals between years. It uses last year’s data to calculate the total area of vacant and derelict expected this year and compares this to actual reported total area this year

GIS checks

These checks are undertaken using mapping software called ARCMAP and are made on the SVDLS-A dataset (new sites) only. The shapefile contains containing the geographic co-ordinates of new sites are mapped. The Excel version of SVLDS-A dataset is read transformed into a shapefile and also mapped. Both maps are overlaid to check that the data submitted in Excel corresponds to the data submitted in the shapefile.

New sites

The following checks are made:

• check all sites lie within the Planning Authority boundary

• check all sites on register have corresponding polygon and vice versa. The sites on the Excel file appear as point data on the map. The sites on the shapefile appear as polygon of the whole site. If there is a point but no polygon, check, and vice versa

• check the area of the site polygon matches site size on returned spreadsheet
• check East/North co-ordinates match nearest position within the polygon to the polygon centroid i.e. the point data should appear at the centre of the polygon and not way offside

Publication checks

In the statistical bulletin each data point is double checked by a team member against the raw data. The same is done for any infographics. All text is also proof read. 

All tables, charts, figures and annexes are proof read. 

The SVDLS register is checked and proof read prior to publication.

All published outputs are then signed-off by the senior member of the team.

The surveys forms are double checked before they are sent out to data providers.

Scottish Government
March 2022

Annex A

Variables in SVDLS-A (new sites)

Name	Description	Valid Values
COUNCODE	Council Code
SITECODE	Unique LA Reference for Site	Up to 20 characters
PREVCODE	Code used in previous survey	Up to 20 characters
DBHIST	Flag to indicate if new or existing site	1=new, 2=existing
NAME	Site Name	Up to 50 characters
ADDRESS1	Site Address 1	Up to 100 characters
ADDRESS2	Site Address 2	Up to 100 characters
ADDRESS3	Site Address 3	Up to 100 characters
SITESIZE	Site Size	To 2 decimal places
SITETYPE	Site Type (Vacant or Derelict)
OWNER1	Owner 1
OWNER2	Owner 2
INSPYY	Year of first Inspection	YYYY e.g. 2021
TIMEVD	Flag to indicate when became Vacant/Derelict
PREVUSE	Previous Use
DERCHAR	Derelict Characteristics
INTRLOCN2	Site Location (Settlement or countryside)
DEVPOT	Developability
EAST	East Coordinate	Must enter both East and North or enter OSGRID
NORTH	North Coordinate	Must enter both East and North or enter OSGRID
OSGRID	OS Grid Reference	Include 100km letters e.g. NT09767461 Leave blank if both East and North entered
Comment	Any comments to help identify reasons for change

Variables in SVDLS-B (re-used sites)

Name	Description	Valid Values
COUNCODE	Council Code	See Table 2
SITECODE	Unique LA Reference for Site	Up to 20 characters
PREVCODE	LA Reference used in previous survey	Up to 20 characters
NAME	Site Name	Up to 50 characters
ADDRESS1	Site Address 1	Up to 100 characters
ADDRESS2	Site Address 2	Up to 100 characters
ADDRESS3	Site Address 3	Up to 100 characters
SITESIZE	Area brought back into use or reclaimed	To 2 decimal places
EAST	East Coordinate	Must enter both East and North or enter OSGRID
NORTH	North Coordinate	Must enter both East and North or enter OSGRID
OSGRID	OS Grid Reference	Leave blank if both East and North entered
INUSEMM	Month brought back into use	MM e.g. 03 for March
INUSEYY	Year brought into use	YYYY e.g. 2017
NEWUSE	New Use
MIXED	Flag to indicate if Mixed New Use	Yes=1, No=2
NLUC	National Land Use Classification
FUND	Funding source used
VDLFUND	Flag to indicate if funding from vacant and derelict land fund	Yes=1, No=2