Developing a strategy for PAI at
British Waterways
General Overview of British Waterways
• 2000 mile, 250 year old network of canals and river navigations from
Somerset to Inverness.
• Leading GI user:
• Part of the Central Government SLA
• Wide user base - 1500+ users
• Desktop, mobile and intranet GIS
• Small “expert” community of 10 GIS specialists, scattered
throughout the organisation.
Assessing the impact on BW
• Polygons covering 14000+ ha of land
• Network representing 3400km of waterway
• Asset data includes 6000 bridges,
1600 locks etc
• 2500 listed structures
Assessing the impact on BW
• 70% of all features affected by PAI
•Timescale of Delivery
• 20% by March 2004
• 28% by end of 2005
• 52% by end of 2006
The Challenge for BW
• GIS is a primary information source
• Operate a multi-editor environment
• Affected throughout the PAI programme
• Data cannot be left until 2007
• Wish to implement MasterMap
The Strategy
Data Audit
• Part of a wider data quality excercise.
• Is the data affected?
•Is the data fit for purpose?
•Should the data be shifted or recaptured?
Data Shift Methodology
• “CONSTRAINED RUBBERSHEETING”
• Datasets constrained by topological rules.
•Ensures that features share edges.
•Topologies affected across datasets
• ESRI PAITools used to shift the data.
– Uses OS link files to affect rubbersheet shift.
• Allows for easier QA checks - important for complex objects and large
datasets.
fig.1. Non-topological data
Nested Poygons with differing numbers of
vertices along a snapped boundary
fig.2a. Topological data
Adjacent polygons share edges
fig.2b. Topological data
Topology acts across datasets
fig.3. Data pre-PAI
fig.4. Data post-PAI
Fig.5a. Shifting the Data
fig.5b. Shifting the Data
fig.5c. Shifting the Data
Planning for Action
• PAI Change management team created.
• Responsible for communication of PAI related information.
– To GIS Specialists
– To GI data users - via GIS Specialists and intranet service
Fig6. GISPortal
- web-enabled GIS service
Planning for Action
• PAI Change management team created.
• Responsible for communication of PAI related information.
– To GIS Specialists
– To GI data users - via GIS Specialists and intranet service
• With a multi-editor environment detailed communication of the PAI
datashift programme is considered critical.
Implementation
• Shift undertaken on frozen snapshot of live data.
• Stages involved:
Segment datashift into block areas (by blocks of dmus)
Communicate segmentation and planned freeze date
Extract data and communicate this action
Carry out PAI datashift process
Load POST PAI OS data
Splice shifted data back into live environment
Identify this action to data custodian for user QA
• Metadata provides information on datashift at both dataset and feature
level.
Conclusions
• Don’t stick your head in the sand.
• Use the opportunity as a chance to review your data – ensure you
plan in QA and change management processes.
• Don’t rush into anything, plan your approach carefully and
communicate widely.
• Don’t hesitate to ask for help