Neural,iA data layer
As an example, Argentina's catalog. It's built the same way in every country, with that country's own official sources.
every synthetic persona comes from data you can trace.
We build every population from three layers: public records that set the baseline, third-party data that adds recent behavior, and what your organization already knows about its own decision.
No source goes in raw or alone. When it agrees with the others, the signal gets stronger. When it disagrees, it gets flagged and weighs less.
Each colored dot is a source in the catalog and opens its details.
explore the sources.
The public baseline is always there. The other two layers come in when the study needs them.
no source gets in alone.
One measure, seen from different sources. When the bands overlap, the signal gets stronger. When one falls outside, it gets flagged.
the limits travel with the data.
What you see above is a sample of the catalog. Each source is stored at the granularity and frequency it really has, without filling gaps to make the map look complete. Each source's coverage gaps, lags and biases go into the model as declared limits and show up in the report.
Comes from an official source, a vendor or the client. Stored with its download date and file hash.
Comes from a model. Carries the model version, the seed and the data it used.
Generated by the synthetic personas. Anything derived inherits the weakest label among its inputs.
Whatever a campaign, a city government or a brand uploads stays isolated. No other client sees it, not even in aggregate.
Charts are illustrative unless marked “Real data” or “Current rules”