Data quality
A valid reading is one that could actually be true — the right unit, a value inside the range the equipment allows, the shape the measurement is supposed to have. When nothing checks that, a wrong number does not look wrong. It just quietly becomes the basis for a decision.
What it means
Validity is the plainest of the six dimensions and the easiest one to skip. It asks a single question of every value — is this the kind of number this measurement is allowed to be?
A pressure below zero, a temperature recorded in the wrong unit, a running-hour count that gained a year overnight. None of these are close calls. They are values the real world does not permit, and they should never have travelled as far as a report.
What it costs
Invalid values rarely announce themselves. They move through reports, dashboards and decisions looking exactly like the good ones, and the cost lands weeks later on someone who had no way of knowing.
A flow rate recorded in one unit and read as another does not look broken; it looks like the process changed. Teams spend days chasing a fault that was never there, and by the time the unit is found, the shift report and the decision built on it are already wrong.
A negative pressure or a temperature far outside anything the equipment can produce still gets averaged, charted and reported like any other number. A single reading of that kind can pull a monthly figure far enough to justify maintenance nobody needed, or to hide a trend that mattered.
Once an operation has been burned by a value that could not have been real, people start verifying figures by hand before they act on them. The platform still runs, but the speed it was bought for is spent on cross-checking.
Our approach
Validity is not something you inspect after the fact. It comes from how a measurement is stored in the first place — with its unit attached, tied to the asset that produced it, and described as a particular kind of measurement rather than an anonymous number.
Every value is stored together with the unit it was measured in, so nothing has to be inferred from a name or from local habit. A number that arrives in a different unit than the rest can be seen for what it is, instead of blending into the series beside it.
Measurements are tied to the asset that produced them and to the kind of thing being measured, not just to a label somebody typed once. A reading is then judged against the equipment it came from rather than against everything else in the plant.
What counts as a plausible value is a property of the equipment and the process, so it belongs next to the asset rather than buried in a report somewhere else. That keeps the limits beside the thing they describe, and a reading that falls outside them stands out against its own context.
The other dimensions
Quality is not a single number. Each dimension fails in its own way, and each one is worth understanding on its own terms.