Widgets and Gauges
Single-value displays that carry status — the fastest thing to read on a dashboard, and the easiest to overuse.
A Widget shows one value large and clear, usually with a status colour. On a dashboard it is the element people read first and the element that carries the most meaning per square centimetre — which is exactly why it needs restraint.
Creating one#
Create a Widget with the constructor for your interval, then drop a measure on the Sections container. The chart type builds the Section, the Widget, and the measure item with its limit, event-attribute, and colour items beneath it.
One Widget chart can hold several sections, so a Widget is not necessarily one number — it can be a small block of related numbers sharing a heading.
Colour is the whole point#
A number alone tells you the value. A number with a colour tells you whether to do something about it, which is a different and more useful thing.
Colour comes from the measure's limits. Configure the thresholds on the measure, bind the limit onto the widget's measure item, and the widget takes its colour from where the current value sits.
Configure the limit, not the colour. It is possible to set a fixed colour on a widget. Resist it, except for genuinely static decoration. A hard-coded colour is a lie waiting to happen — it will stay green while the value goes out of specification. Drive colour from limits and it stays true without maintenance.
Keep the palette conventional and consistent across the whole dashboard set: red for out of specification, amber for marginal, green for good, grey for no data. Inventing a scheme per dashboard means every reader has to relearn it.
Text instead of numbers#
Because a widget displays the measure's formatted value, the three-section format string turns a numeric measure into a word:
| Format | Positive | Negative | Zero |
|---|---|---|---|
'Pass';'Fail';'Fail' |
Pass | Fail | Fail |
'Running';'Fault';'Stopped' |
Running | Fault | Stopped |
'On target';'Below';'On target' |
On target | Below | On target |
A calculated measure returning 1 or 0, formatted this way, with a limit for colour, gives you a status indicator with no special chart configuration at all. This is the cheapest useful thing on an Infohub dashboard and it is under-used.
What belongs on a widget#
A widget is for a value where the current number is the point and its trend is secondary.
Good candidates:
- Today's production against target
- Current OEE or availability
- Month-to-date consumption
- A pass/fail quality status
- Time since the last incident
- A count of open stoppages
Poor candidates:
- Anything where the shape over time is what matters — that is a trend chart
- A value that changes so fast the number is never stable
- A number requiring context to interpret, which the widget has no room to give
The test: if a reader's next question is always "and how does that compare with yesterday", the widget is the wrong element and a small trend chart is right.
Gauges, bullets, and the alternatives#
Infohub has more than one way to show a single value against a target.
Widget — the number, large, with status colour. Highest information density, works at any size, readable across a control room.
Bullet — a bar against a target marker with qualitative bands behind it. Shows the value, the target, and the range in one compact horizontal element. Genuinely good for actual-versus-target rows, and more space-efficient than a gauge.
Radial gauge styling — familiar to operators from SCADA, and honest about a value's position within a range. Costs several times the space of a widget for the same information, which is the reason to be sparing.
For a column of ten KPIs against target, a set of bullets communicates more in less space than ten gauges and is easier to scan than ten bare numbers.
Gauges are the most over-used dashboard element. A screen of twelve gauges looks impressive in a screenshot and is hard to read in practice — the eye has no natural scan order, and each gauge is mostly empty space. Use gauges for the two or three values that genuinely warrant the emphasis, and widgets or bullets for the rest.
Sizing on a dashboard#
Dashboards use a 12-column grid. A widget sits comfortably in about 3 columns by 3 rows, which gives four across a row.
Size to viewing distance, which is the consideration people forget. A control room screen read from four metres away needs far fewer, far larger widgets than a desktop dashboard read from arm's length. Building one dashboard for both audiences satisfies neither. See Dashboard Layout Principles.
The interval question#
A widget shows the value for the current period at its interval, so the interval determines what the number means.
A Daily widget shows today — which is a partial day until midnight. That is usually intended for a "today so far" number, and it is a genuine source of confusion when someone compares it against yesterday's complete figure and concludes performance has collapsed.
Where the partial-period reading is likely to mislead, either label it explicitly — Production (today so far) — or show the last complete period instead.
Diagnosing#
| Symptom | Cause |
|---|---|
| Widget colour never changes | No limit bound, or a fixed colour was set |
| Colour is wrong | Limit thresholds, or the limit belongs to a different measure |
| Shows no value | Measure has no value for the current period, or the interval is wrong |
| Number looks far too low near the start of a period | Partial period — expected; label it or show the last complete one |
| Shows full floating-point precision | No format string on the measure |
| Text format shows a number | Format string not using the three-section pattern |
| Unreadable at distance | Too many widgets, each too small for the viewing distance |