How the market snapshot is assembled
One bidding zone, one delivery day, pulled from the source APIs at the moment you ask for it. Every other view in the console reads what was collected overnight; this one does not, and that difference decides what it is good for.
What the screen answers
What happened in one zone on one delivery day, and what physically explains it: the published price and its shape, the markets around it, and the generation and load behind it.
It is the orientation screen. An analyst opens it to understand a day before deciding whether any hour in it is worth reconstructing.
Live at request time, not from the archive
Each panel calls the publisher when the page loads. That is why the day can be today, why the price updates as the auction clears, and why the page is slower than the stored views. It is also why a rate-limited or failing feed degrades its own panel and reports itself in the sources table, rather than emptying the page.
The stored archive appears once, at the bottom of the price section, as context for where the day sits against the history we hold. It is never the source of anything above it. A number on this page and the same number on a stored view can differ, and when they do the stored one is a snapshot and this one is now.
The controls
- Zone
- The bidding zone whose delivery-day clock everything is cut on. Change it and every panel re-reads.
- Compare
- Up to a fixed maximum of further zones, drawn against the first. They are cut on the FIRST zone's clock so the columns can be read against each other, rather than each zone being described in its own terms.
- Delivery day
- Today by default, which is the only setting where the Live badge appears. Tomorrow is available once its auction has cleared.
- Currency
- Whether to convert onto one euro axis, and at which rate. GB publishes in sterling; nothing is converted at a guessed rate, and a converted chart prints the rate, its source and its date - including when the delivery day was a weekend and the last fixing before it was used.
The four sections
- Price
- The headline numbers, the curve they come from, and where the day sits against the stored history for the zone. The tiles carry price now, day average with a sparkline, the next day or the imbalance price, the range, the peak/off-peak spread and the steepest ramp.
- Compared markets
- Every selected market on the shared clock, then the hour-by-hour view of where they came apart: a heatmap of the levels and diverging bars of the spreads, with a table of the same numbers underneath.
- System
- The physical picture behind the price - generation mix, residual load, renewable share, the day-ahead forecast against what actually generated, and what crossed the borders.
- Sources
- Which call answered, how fast, and what did not. It is the last section because it has to wait for every other one: what answered and what did not is only true once every call has come back.
What the sources table is telling you
Every panel names the feed that answered it by the ROLE that feed plays, never by who publishes it: a regulatory publication, an exchange's own results, a secondary reference compilation, or a published FX fixing. That distinction is the one that changes how much weight a number carries.
A failed call is reported as failed with its reason, and a call that succeeded and returned nothing is reported as nothing published - not as success. Those are different facts about the market and the page never collapses them.