A data pipeline is a series of steps that takes large data sets and creates usable results from them. At the beginning of a data pipeline, a data set might be pulled from a database, a distributed file system, or a Kafka topic. Throughout a data pipeline, different data sets are joined, filtered, and statistically