Skip to content

Data Lineage

Complete traceability from source data to model parameters

VerveStacks maintains complete data lineage documentation for all model parameters. Instead of every modelling team independently cleaning the same global datasets and inventing its own mapping rules, the platform turns that once-off “ground-clearing” into a shared asset that can be reused across countries, studies, and frameworks. This reduces duplication while making it easier for analysts to see exactly how any number in a country model was derived.

Lineage Documentation

Source Tracking

  • Every model parameter traced to original data source
  • Transformation steps documented with code references
  • Data vintage and update information maintained
  • Quality flags and validation results included

Processing Documentation

  • Data cleaning and validation procedures
  • Gap-filling methodologies and assumptions
  • Aggregation and clustering algorithms
  • Parameter derivation formulas

Output Documentation

  • Model parameter derivation explanations
  • Assumption documentation and rationale
  • Uncertainty and sensitivity information
  • Validation against independent sources

Implementation

Data lineage is automatically generated during model creation and included in model documentation packages.

Note

Interactive data lineage visualization tools to be implemented.