PAI Toolbox: Data audit

In order to plan for PAI it will be necessary to analyse the current situation and determine the relevant parameters. Since Positional Accuracy Improvement (PAI) instantaneously affects geospatial user data, an overview of the data sets used in current business processes will be required. You will be lucky if you already have a detailed and maintained data inventory-if not, this is the perfect opportunity to create one. After that, the PAI impact can be assessed on every individual dataset and prioritised. Another option would be to list all relevant business processes, link them to the datasets and prioritise the datasets in relation to the priority of the business process.

The reducing the impact of positional accuracy improvement document gives a few more ideas and introduces further ideas to extend the audit to users and data sharing procedures and GPS data capture.

The data audit matrix provides a template to help you assess your own data, how it was captured and how it is being used. This will aid in the decision-making process on what you should be doing with the data. The data audit matrix is taken from the document audit questionnaire to assess the impact of PAI on a specific dataset.

Example: Stratford-on-Avon District Council

Stratford-on-Avon District Council trace/copy Land-Line^® for their area/polygon data, ready for NLIS Level 3 and it was always better for collecting "data within polygon" enquiries on the GIS. This meant that rubber sheeting was never an option for our datasets captured to this accuracy.

The following tables illustrates the different types of data sets held by SDC and the specific PAI action for each.

Type	Category	PAI Action
a	Those captured at 1:10 000 to 1:50 000 3^rd party data, most utilities, Environment Agency, English Nature data, some OS datasets (Boundary-Line^®)	Visual check; for falling wrong side of obvious landmarks (buildings, water, roads). Communicate viewing scale accuracy (e.g. on GIS legend)
b ~	In-house Land-Line data captured datasets (including recapture of some 1:10 000 supplied datasets).	Identify datasets Prioritise ‘Business Critical’ datasets Identify numbers of features in PAI area for each dataset*
c ~	As b but with freeform lines not associated with Land-Line or a measured distance from Land-Line. (same for OS MasterMap^®)	Identify datasets Prioritise ‘Business Critical’ datasets Identify numbers of features in PAI area for each dataset*
d ~	MS Access etc databases viewed in/ linked to GIS	Identify datasets Prioritise ‘Business Critical’ datasets Identify numbers of features in PAI area for each dataset*
e	Datasets about to be data captured	Areas OK to do (1:1250 and PAI areas). Deadline?
f	“Don’t start” – datasets	Assess if pilot will do for now, see e
g	Base data – OS, Aerials	Still needs PAI management
h	Project files – date stamp eg Summer Leisure schemes, landscape assessments	Question still relevant, access scales for a,b,c; store as archive against archived Land-Line for year
j	Datasets that you provide to other organisations/authorities (i.e. we are the third party)	What have they done with PAI? When do they need PAI datasets to match their base map holdings? Do you need to keep two parallel versions?
k	Datestamp/Vintage of National Initiative maps eg. NLIS hub, NLPG grid references	Has anyone got an answer on this yet?

* Draw PAI tiles as a dataset, perform GIS analysis for any part of selected="selected" dataset features falling within PAI tiles = number of features requiring attention.

~ Large polygon/areas and those overlapping PAI area are not likely to fit solutions

Metadata is never fun to collect, but is invaluable in informing you about your dataset holdings, and as we already held this metadata it proved to be a fast starting point in identifying datasets and priorities. Examples of metadata we hold are given below.

Stratford-On-Avon PAI metadata

For more information see the Stratford-On-Avon case study.

Angus Council - Hamish Thomson

The most important stage in dealing with the issue of PAI is to do a proper assessment of your data, to find out what you have, how was it created, what scale of mapping was used, what system it is held in and whether it was snapped to the base map. It is an ideal opportunity to create or update metadata in your organisation.
The viewing scale also needs to be assessed - if at a large scale (1:1250 / 1:2500) then it will be affected by PAI, even if it was created at a small scale.
If the data is used in systems where staff depend on it on a day to day basis or where automatic constraint checking is used then the impact of PAI may be more severe.
When assessing the impact of PAI systems need to be analysed to find out what GI data is used in them, how is it stored, how it can be extracted, moved, quality assured and put back in again and who will do each of these stages.
By taking all of these factors into consideration it is possible to analyse the results and produce a prioritised listing of data that will be affected by PAI and what should be moved and in what order, along with any costs of managing data in systems.

Back to Find a sponsor Forward to Revised risk assesment