Real-estate data often arrives in compact form: a rent, a price change, a vacancy rate, a transaction count. The number is easy to display. The harder part is preserving what the number actually describes.
SAID treats an observation as a bundle rather than a scalar. The value travels with its unit, asset class, geography, observation period, publication date, source definition, retrieval date and transformation status. Removing those attributes may make a dashboard cleaner, but it can also change the meaning of the evidence.
Why lineage matters
Two values can look identical while answering different questions. A monthly rental index can show the direction of rents without telling you the average rent paid. A registered transaction count can show administrative activity without matching the date when buyers and sellers originally agreed to the transaction. A forecast can describe expected supply without being an observed completion.
Lineage lets the user trace the displayed value back to the institution and dataset that defined it. It also makes later corrections possible. If a source revises a series, a system that retained the release and observation context can distinguish a revised historical value from a newly observed period.
Direct, transformed and derived
SAID separates three broad evidence states. A direct observation is displayed substantially as the official source reports it. A transformed observation uses a documented operation such as rebasing, unit conversion or percentage-change calculation. A derived observation combines inputs or applies a method that creates a new analytical value.
The distinction is not a hierarchy of quality. Transformation can improve comparability, and derivation can answer useful questions. The requirement is that the method remain visible enough for the user to understand what changed between source and display.
What to inspect before using a number
- What exactly is being measured?
- Which properties or housing segment are included?
- Where does the observation apply?
- Which period does the value describe?
- When was it published and retrieved?
- Is the displayed value direct, transformed or derived?
- What limitation would materially change interpretation?
This discipline is deliberately conservative. It reduces the number of apparently comparable values, but it increases the amount of usable evidence.