Qanat
Water and power, metered in real time
Client identities and brand marks are anonymised. Names, logos and interfaces shown are not real companies or real projects.
- Sector
- Utilities
- Location
- Muscat, Oman
- Year
- 2024
- Kickoff to production
- 23 weeks
A metering platform that ingests 1.2 million readings a day and finds a leak in hours instead of a billing cycle.
Readings were collected monthly by van. Leaks were discovered when a resident complained or a bill tripled, and the network operator had no way to distinguish a burst pipe from a hot summer.
Ingest first, model second
A pipeline built for millions of daily readings with gap handling and back-dated corrections, because real meters go offline and lie.
Anomalies against a baseline
Consumption is compared to each meter's own seasonal pattern, so an alert means something changed rather than that it is August.
Consumption people understand
A resident portal that shows usage in cost, compares it to similar homes and explains a spike in plain language.
Non-revenue water fell meaningfully in the first year, and the field team is dispatched by evidence instead of by complaint.
What changed
- Readings ingested daily
- 1.2M
- Leak detection
- Hours
- Field dispatch
- Evidence-led
- Go
- TimescaleDB
- Kafka
- Next.js
- Grafana
Tell us what's slowing you down.
Fifteen questions, one screen at a time, and a verdict at the end. No call needed to get it.