The Live Feed of Pump Data
Leak Detection on Historical Data
Leaks are often a problem that go unnoticed until the damage has been done. This often results in a large waste of water, damages to property and increased bills for residents and owners of the property.
What it does
Grundfos DataViz queries Grundfos' pumps for live data about the water flow and analyzes historical data to try detect outliers that represent abnormal water usage. The hope is then to use this to quickly detect possible leaks.
How we built it
We hacked, chopped and stitched it together using absurd amounts of coffee and glue.
Challenges I ran into
Classic Hackathon compatibility issues and sleep deprivation.
Accomplishments that I'm proud of
Never used Django before, never did outlier detection
What I learned
Bits and pieces about Django, Charts.js, Outlier Detection
What's next for Grundfos DataViz
Next comes machine learning on the historical data to try to predict leaks in real-time on the live data. Notifications through i.e. email, push notifications, etc., and in general a lot of UX optimizations.