SNAP Data Catalog

The Data Catalog is the central entry point for discovering datasets in SNAP.

It supports thematic browsing, faceted filtering, and metadata-first selection so users can quickly determine whether a dataset is fit for purpose.

Catalog structure

Datasets are grouped by domains:

• Jobs & Employment
• Skills & Training
• Demographics
• Socio-economic Conditions
• Policy Signals & Context

Each dataset card includes:

• Title and short description
• Data owner / source institution
• Geographic granularity (EU / country / NUTS1 / NUTS2 / local)
• Temporal coverage
• Last refresh timestamp
• Quality badge (simulated)

Search behavior

For debugging, search should match:

• Dataset title
• Dataset description
• Tags
• Source name
• Indicator codes

Example terms to test:

• labour mobility
• vacancy
• female employment
• NUTS2

Filters

Use the following filters in combination:

• Country / Region
• Reference period
• Theme
• Data type (time series, cross-section, microdata)
• Update frequency (monthly, quarterly, annual)

Quality and trust indicators

SNAP should display quality dimensions:

• Completeness: percentage of expected records present.
• Consistency: proportion passing schema and range checks.
• Timeliness: delay between source publication and ingestion.
• Documentation score: metadata richness and clarity.

Comparison mode

Comparison mode lets users evaluate two datasets side by side:

• Coverage by geography and years
• Unit and indicator definitions
• Refresh cadence
• Known caveats

Use this mode to avoid mixing incompatible indicators in one analysis.

Saved catalog views

Users can save a filtered catalog state:

• Name: snap-catalog-skills-eu-annual
• Shared with: team or private
• Optional note: use-case explanation

Debug scenarios

1. Search for skills mismatch and apply annual filter.
2. Open three datasets and compare coverage.
3. Save the view and reload it.
4. Confirm all filters are restored.

Troubleshooting hints

• Empty result sets usually come from over-filtering.
• Missing last refresh likely means incomplete metadata payload.
• Slow search often indicates a non-indexed text field.