Smart energy meter readout portal

Smart energy meters in the Netherlands deliver a continuous stream of measurement data through the so-called P1 standard — consumption, feed-in, gas, phase load. In itself this is a treasure of information, but only useful if you can reliably collect, store, and visualize that data in multiple ways.

For this project I worked on a site that does exactly that. Goal: collect measurement data from smart meters and display it in various ways — for end-users who want to understand their consumption, and for parties who want to see patterns at larger scale.

My contribution mainly ran on two tracks: designing a repeatable way to import measurement data, and setting up an effective system to detect anomalies. Both sound like plumbing, but they are crucial: without reliable ingestion every downstream visualization is misleading, and without good anomaly detection you only notice weeks later that a whole stream of data has been leaking.

For the import I had to learn a lot about the P1 standard. On paper it looks tidy; in practice there is variation between devices and firmware versions that the spec doesn't make obvious. We therefore built the parser defensively: better to keep an unknown field clearly marked and pass it along to analysis than to silently throw it away.

Anomaly detection was built into multiple layers: missing batches, values that diverge from the historical pattern, sudden jumps that are physically improbable. Not to intervene automatically — the world is too messy for that — but to raise an early, clear flag with operations and the data team.

Technically we worked with Silverstripe and PHP for the web side, with jQuery for light interactivity. Nothing complicated; precisely by combining simple tools with a thought-through data architecture we could efficiently deliver what the client needed. Proud of the way we got this implemented.