Road Damage Reporting with a Live Dashboard & Automated Dispatch
Citizens and field crews report damaged roads in seconds; reports land on a live map dashboard and trigger automated routing to the right maintenance team.
- Duration
- 10 weeks
- Services
- Custom SoftwareWorkflow AutomationAI Integration
A regional infrastructure unit collected road-damage complaints through phone calls, chat messages, and paper forms. Reports were logged into a spreadsheet days later, locations were vague, and photos lived in someone's phone. Management had no map of where the worst problems were, and crews were dispatched by whoever shouted loudest rather than by severity.
Challenges
- Reports arrived through scattered channels with no location or photo standard.
- Duplicate complaints about the same pothole were logged as separate issues.
- No one could see, at a glance, where damage was concentrated.
- Dispatch was manual and reactive, with no link between a report and its fix.
Objectives
- One simple way to report damage, with photo and precise location.
- Automatically classify and de-duplicate incoming reports.
- A live map dashboard showing severity, status, and hotspots.
- Automated routing of each report to the responsible maintenance team.
We built a mobile-friendly reporting form that captures a photo and GPS location, then a pipeline that classifies damage severity from the image, clusters nearby reports so duplicates merge, and plots everything on a live dashboard. Workflows assign each cluster to the correct team by zone and severity, notify them, and track the report through to a verified repair.
Implementation
- 01Reporting flow & taxonomy
Two weeks designing the citizen and crew reporting flows and agreeing a severity taxonomy with the maintenance teams.
- 02Classification & de-duplication
An image-classification step scores severity; a geospatial clustering step merges nearby reports into single work items.
- 03Dashboard & dispatch automation
A live map dashboard with status and hotspots, plus workflows that route work items to teams by zone and notify them.
- 04Pilot & rollout
A one-district pilot for three weeks to tune the severity model and clustering radius, then rollout with team training.
Architecture overview
- Next.js reporting form with photo capture and GPS
- Image-classification model for severity scoring
- PostGIS-backed clustering for de-duplication and hotspots
- n8n workflows for zone-based routing and notifications
- Live map dashboard with status tracking through to repair
What changed for the business.
- Report-to-logged time went from days to the moment of submission.
- Duplicate complaints collapsed into single, trackable work items.
- Management can see damage hotspots on a map instead of a spreadsheet.
- Dispatch is driven by severity and zone, not by who called first.
Lessons learned
- A GPS-and-photo form removes the ambiguity that made old reports unusable.
- De-duplication is what makes the dashboard trustworthy. Without it, one pothole looks like ten.
- Closing the loop to a verified repair matters as much as capturing the report.
Facing a similar problem?
Tell us about your operation and we'll show you what the same delivery method would look like for you.