A store visit leaves a record of something that happened. A forecast of next month’s visits is different: it is new data created from earlier records. That is derived data—an extension of raw, naturally occurring data produced through analysis, often by algorithms that carry observed patterns forward. section 7.2.3
Forecasting store activity
Store visits follow weekly, seasonal, and holiday patterns. Before a predicted storm in the US Midwest, shoppers may buy more milk and bread. Analysts can combine past visits with information about store hours, population, and other conditions to estimate future visits and sales. Those estimates can then inform forecasts for staffing, stock levels, and promotions. The forecasts are newly created data, not records of visits or purchases that have already occurred. section 7.2.3
A record does not stop being useful after its first analysis. Data can be aggregated, combined with other sources, and extended into further data. Its lineage connects those later products to the records used to make them. Derived data can also feed synthetic data: estimates for proposed stores, for example, can draw on both raw and derived information when no store activity exists yet to measure. section 7.2 section 7.2.4 section 7.7
Why does that connection matter to you? Your visits and purchases can contribute to forecasts that serve purposes beyond the original transaction. Tracing that path helps make the new data—and questions about who benefits from it—visible. section 7.0 section 7.2.3
