Skip to main content

Data Methods Overviews

This section provides a unified view of how Carbon Arc applies logic, scores data quality, and constructs reliable models across its panel-based datasets.

At Carbon Arc, methodology is a multiplier. Our goal is not just to provide data, but to encode confidence, clarity, and consistency. This section documents the operational logic behind the platform — so users can understand how our data is processed, filtered, and interpreted.

What’s Included​

We break methodology into two crtical domains:

Data Logic​

Outlines the transformation logic applied to clean and panelize raw data across domains. These pages document decisions made to improve accuracy, structure, and comparability.

Data Models​

Defines how we evaluate dataset confidence, user behavior, and derived metrics. These models turn raw transactions into insight-ready dimensions.

  • Confidence Scoring: Quantifies how reliable each panel is across volume and trend fit
  • Growth Metrics: Measures directional accuracy and trend alignment
  • Overlapping Shoppers: Identifies shared customers across brands using statistical overlap
  • Loyal Shopper: Classifies and scores user loyalty monthly with cadence + quantity logic

How to Use This Section​

Use this section to:

  • Understand the logic behind any panel’s construction
  • Interpret metric creation logic
  • Track how filters and models evolve over time
  • Align analysis and reporting to certified methodology

Our methodology section is designed to eliminate ambiguity and empower confident usage of our data. Please reach out to support@carbonarc.co with any questions!