GIS Field Data Collection and Validation Middleware
Challenge
Utility and roadway field crews capture asset data with mobile apps, often without connectivity, and that data has to enter a geographic information system that acts as the record of truth. Writing straight from the device was corrupting records: submissions arrived incomplete, with free text in coded fields, out of range measurements, or mixed datum realisations that shift buried assets. Offline edits created conflicts that needed a defined policy, and positional accuracy metadata missed at capture could never be recovered. Pipeline, locate and roadway records carry legal weight, so rejected submissions, reviewer decisions and rule versions all had to be retained as compliance evidence.
Utility and roadway field crews capture asset data with mobile apps, often without connectivity, and that data has to enter a geographic information system that acts as the record of truth. Writing straight from the device was corrupting records: submissions arrived incomplete, with free text in coded fields, out of range measurements, or mixed datum realisations that shift buried assets. Offline edits created conflicts that needed a defined policy, and positional accuracy metadata missed at capture could never be recovered. Pipeline, locate and roadway records carry legal weight, so rejected submissions, reviewer decisions and rule versions all had to be retained as compliance evidence.
Solution
We built middleware between the capture applications and the record system. Capture apps work on scoped offline replicas and synchronise with a defined conflict resolution policy, recording receiver type, correction source, accuracy estimate and coordinate reference with every point and attaching photos and forms with full provenance. A declarative, versioned rule engine validates required attributes, coded domains, cross field dependencies, ranges, snapping, connectivity, duplicates and utility network topology, with configurable severity so hard failures block the load and soft failures go to review. Validation runs at submission so crews get feedback on site. Accepted data is loaded transactionally through versioned edit sessions, and geospatial analytics and supervisor dashboards run on top.
We built middleware between the capture applications and the record system. Capture apps work on scoped offline replicas and synchronise with a defined conflict resolution policy, recording receiver type, correction source, accuracy estimate and coordinate reference with every point and attaching photos and forms with full provenance. A declarative, versioned rule engine validates required attributes, coded domains, cross field dependencies, ranges, snapping, connectivity, duplicates and utility network topology, with configurable severity so hard failures block the load and soft failures go to review. Validation runs at submission so crews get feedback on site. Accepted data is loaded transactionally through versioned edit sessions, and geospatial analytics and supervisor dashboards run on top.
Results
Crews now learn about defective submissions while they are still next to the asset rather than days later. The record system receives only conforming data with coordinate reference and accuracy metadata intact. Regulatory evidence is complete because rejected submissions, reviewer decisions and rule versions are kept alongside accepted records. Supervisors see backlog, submission quality and crew throughput, and analysts have data clean enough for defect clustering, inspection route prioritisation and failure likelihood modelling.
Crews now learn about defective submissions while they are still next to the asset rather than days later. The record system receives only conforming data with coordinate reference and accuracy metadata intact. Regulatory evidence is complete because rejected submissions, reviewer decisions and rule versions are kept alongside accepted records. Supervisors see backlog, submission quality and crew throughput, and analysts have data clean enough for defect clustering, inspection route prioritisation and failure likelihood modelling.