Where did this data come from, what changed it, and where is it being used? Provenance—also called data lineage—answers those questions by tracing a data item back to the original data and recording the steps that created it. section 7.1
Data rarely stays in its original form. Raw activity can become aggregated, integrated, derived, synthetic, simulated, or optimization data. Each new asset remains connected to the data used to create it, even when analytics produces something substantially new. section 7.0 section 7.7
Tracking provenance makes data use more accountable:
- Identify which people, organizations, or systems contributed to a blended dataset. section 7.1
- Connect permissions, monetization, and compliance obligations to the data that supports a new asset. section 7.1
- Audit where data has been used and track payments when the same source supports multiple applications. section 7.1
- Manage and protect data as it moves beyond the activity that originally produced it. section 7.0
Purchase data becomes multiple analytical products
A retailer licenses a person’s purchase data to study personal-care buying, then blends it with household-cleaning purchase and use data. Provenance records both uses, the millions of contributing sources, and the resulting payment and audit obligations. section 7.1
Forecasts retain a connection to everyday activity
Store visits, economic conditions, crime, store expansion, hours, and population rates can produce forecasts for sales, staffing, stocking, shortages, and promotion effectiveness. Those derived streams retain lineage to visits and purchases. section 7.2.3
