Publish a small L0 Populace dataset for web app and Colab

Context

The full Populace replacement for legacy policyengine-us-data needs a smaller companion artifact for interactive use in the web app, examples, and Colab notebooks.

Goal

Produce and publish a small Populace US dataset around 20k records using L0 sparsification, derived from the same validated full-release pipeline rather than a separate hand-built sample.

Acceptance criteria

• Starts from a certified full Populace release artifact.
• Uses an L0/sparsification step that preserves calibrated fiscal/geographic behavior as well as possible for its size.
• Exposes diagnostics against the standard current target surface, including congressional-district targets, with clear comparison to the full release.
• Is small enough for web app and Colab use.
• Has an explicit release manifest entry and documented intended use, so it is not confused with the full production replacement dataset.

Notes

This should not block the big-dataset policyengine-us-data replacement, but it should be tracked as part of making Populace usable across production, web app, and notebook contexts.

Start by locating the validated full Populace release pipeline, its certified artifact, and the existing L0 or sparsification entry point. Define the small dataset from that pipeline, compare diagnostics against the standard target surface and full release, then add the release manifest entry and intended-use documentation for web app and Colab use.