Technical detail ยท Cloud Data: Quality and Security
Smart-meter minimization and release evidence
The meter fleet illustrates data minimization: keep the smallest useful detail for each purpose. Billing, grid balancing, and customer support need different products, roles, and retention policies. Provenance identifies the origin of each reading; lineage connects it to the product.
Worked example: smart-meter minimization fleet: 3,000,000 smart meters reporting: one reading every 15 minutes readings per meter per day: 96 daily readings: 3,000,000 * 96 = 288,000,000 readings/day privacy-aware data products: billing view: meter-level kWh by billing period, restricted to billing roles grid-balancing view: postcode or feeder aggregate by hour, no direct customer id customer portal view: customer-specific detail only for authenticated customer and approved support role release evidence: data classification tag purpose for collection retention window access policy id key id aggregation rule lineage from raw reading to aggregate audit log of reads and exports design reading: Minimization is not the same as deleting all data. The pipeline creates separate data products for separate purposes, each with the smallest useful detail level and its own retention and access policy.