A provincial irrigation department managing thousands of canal structures across several operating divisions was still tracking asset condition and maintenance work through paper forms and disconnected Excel workbooks, with three incompatible ID schemes in use across divisions. Work orders routinely took weeks to route between field inspectors and divisional engineers, and nobody had a single authoritative view of where the network's assets actually were.
GeoVeris built a Next.js and PostgreSQL/PostGIS web-GIS portal that digitizes and geolocates the department's canal infrastructure — regulators, escapes, and cross-drainage works — from a mix of as-built drawings and a fresh GNSS-RTK field survey, and gives field inspectors an offline-capable mobile form for logging condition assessments that sync automatically once connectivity returns.
Migrating six years of historical maintenance records into the new system meant reconciling three incompatible legacy ID schemes against the newly surveyed asset locations. We built a matching pipeline that combined fuzzy name matching with spatial proximity scoring, routing only the genuinely ambiguous matches — about 12% of records — to divisional staff for manual confirmation, rather than asking engineers to reconcile the full dataset by hand.
The portal is now the department's system of record for asset condition tracking, and replacing paper-based work order routing with the portal's built-in workflow cut average maintenance turnaround from 19 days to 7.
Methodology
Developed a Next.js and PostgreSQL/PostGIS web application digitizing the department's canal network, control structures, and 3,100+ individual assets (regulators, escapes, cross-drainage works) from a mix of scanned as-built drawings and a fresh GNSS-RTK field survey. Built role-based workflows for field inspectors to log condition assessments and photos from mobile devices (offline-capable form submission with background sync), feeding a maintenance-prioritization dashboard for divisional engineers built with server-rendered React components and chart visualizations. Migrated 6 years of historical maintenance records from disparate Excel workbooks into the normalized PostgreSQL schema, with a validation pass reconciling asset IDs across the old and new systems.
Results & Outcomes
- Digitized and geolocated 3,100+ canal infrastructure assets into a single authoritative GIS-backed database
- Reduced average maintenance work-order turnaround from 19 days to 7 days via workflow-based routing
- Migrated and reconciled 6 years of historical maintenance records with zero data loss reported post-cutover
- Adopted as the department's system of record for asset condition tracking across 4 operating divisions
Challenges
The department's legacy asset records used three incompatible ID schemes across different divisions, none of which matched cleanly to the new GNSS-surveyed asset locations. We built a semi-automated matching pipeline combining fuzzy name matching with spatial proximity scoring against the surveyed coordinates, then routed ambiguous matches (about 12% of records) to divisional staff for manual confirmation through a lightweight review UI built specifically for the migration, rather than asking field engineers to work from raw spreadsheets.