A data pipeline is a structured flow that supplies data to models, analytics operations, and operational engines. Firms depend on sufficient data volume, variety, and velocity to model and simulate the real world. section 2.5.1
Pipelines are only as effective as the data entering them. Seasonal patterns and ordinary behavior changes can be incorporated, but abrupt disruption may require models to be retrained, recalibrated, or rebuilt. section 2.4.7
The pipeline’s dependence on ongoing data creates practical consequences:
- A sudden change such as the COVID pandemic can make established predictions unreliable. section 2.4.7
- Collectively withholding data can interrupt pipelines and alter how companies predict, message, and engage with people. section 2.4.7 section 2.5.1
When the data stops
Platforms that use data to predict behavior around social media, news, shopping, or buy-now-pay-later services cannot produce the same analytical results when the required data is no longer available. section 2.4.7 section 2.5.1
