Methodology
How Datacriterion data packs are built
Use this page to review the source, transformation, validation and delivery logic behind each official-source pack before you buy.
Source layer
- Official sources are preferred whenever the product family depends on public releases.
- Coverage windows and source boundaries are defined before packaging.
- Each pack should be traceable back to its source system and refresh cadence.
Transformation layer
- Raw files are normalized into consistent structures, names and datatypes.
- Where useful, delivery includes both CSV and Parquet outputs.
- Documentation should make the transformed structure readable without reverse engineering.
Current source systems in the catalog
- U.S. Census Bureau Monthly Retail Trade.
- U.S. Census Bureau New Residential Construction.
- U.S. Census Bureau New Residential Sales.
- Eurostat monthly datasets for inflation, labor, industrial production and retail volume.
Standard delivery layer
- CSV and Parquet outputs where that makes reuse easier.
- A sample CSV for evaluation and structure inspection.
- Series dictionary, reproducible notebook and preview charts.
- Coverage report, validation summary and technical README.
Validation and delivery
- Each pack includes clear delivery files and documentation.
- Representative samples let buyers inspect structure before purchase.
- Licensing and allowed reuse stay visible before checkout.
What buyers should not have to guess
- Which sources and coverage windows define the pack.
- How the packaged files are structured and documented.
- Which license tier fits the planned reuse.
If you already know which pack family fits your workflow, go back to a flagship or sample page. If you still need orientation, use the evaluation guide and FAQ before contacting us.