stevens 
A macro dashboard can put employment, inflation, and production on the same screen without making them directly comparable. The series may use different units, describe different periods, and become available on different schedules. A tidy chart can hide those differences rather than resolve them.
Iva Buric, an analyst on our Macro team, built an indicator panel around that problem. I asked Iva to make every displayed signal answer three questions: what is being measured, what transformation produced the value, and what information was available on the selected date? Iva owned the transformation pipeline and the inspection view behind the panel.
Define the meaning before the scale
Iva began with a small set of indicators and a configuration for each one. The configuration describes the source units, the intended transformation, and the direction in which the resulting signal should be interpreted.
That last field is easy to overlook. A higher standardized value does not mean the same thing for every series. We kept the underlying economic interpretation visible instead of treating every positive number as a generic sign of strength.
The first version also separates levels from changes. Standardizing a level and standardizing its recent movement are different experiments. Choosing between them belongs in a reviewed configuration, not in an unnamed calculation inside a chart.
Normalize using the history available
To compare transformed series, Iva added a rolling standardization step. Each observation is evaluated against a defined historical window, with the window length and minimum coverage recorded in the configuration.
The calculation does not use observations from later in the series to normalize an earlier point. That makes the historical display consistent with the experiment we want to run, rather than letting the full dataset determine what counted as unusual in the past.
She also kept the unscaled and transformed values available in the drilldown. A standardized number is easier to compare, but it should not become a replacement for understanding the original measurement.
Keep revisions and freshness visible
Economic history can change as observations are revised. The St. Louis Fed's real-time-period documentation distinguishes information available today from information known at a historical point. Iva used that distinction when organizing the panel's data snapshots.
Each displayed value retains its availability context and a freshness indicator. A series that has not released a new observation should not look as though it supplied fresh information simply because the dashboard refreshed.
We chose to show missing observations as missing, with coverage reported alongside any combined view. Filling a gap with zero would turn an absence of information into an apparently ordinary standardized observation.
Explain the aggregate from its parts
Iva's comparison view keeps the individual series beside any aggregate. A reviewer can see whether a change comes from several indicators moving together, one unusually large observation, or a difference in the available inputs.
The tests reflect those distinctions. An aggregate must reconcile to its configured inputs. Changing the order of the series must not change the result. Adding future observations must not alter earlier transformation windows, and a constant series must trigger the configured zero-variation handling rather than an undefined score.
What Iva owned
Iva owned the series configuration, transformation rules, coverage checks, and dashboard drilldown. She also wrote the documentation explaining how each displayed signal relates to its original data.
The project gives the Macro team a more useful starting point for discussion. The panel does not decide which economic story is right. It makes the measurements behind competing stories easier to compare, inspect, and question.