A pipeline, in generic terms, refers to the series of steps via which a particular data or input passes through to get processed into the final output. Data analysis pipeline also follows the same definition as it involves all steps starting from pre-processing of data to cleaning of data; and extends till the intended visualization is made, the insights are generated and are communicated to the business stakeholders. The main objective of this is to ease out the entire process; and also make the process more readable and manageable from an implementation perspective.
Generically speaking a pipeline has inputs go through a number of processing steps chained together in some way to produce some sort of output.
A data analysis pipeline is a pipeline for data analysis.
Usually they’re done in some graphical environment such as Alteryx or KNIME (with scripting steps in R or Python, say), each step logically following each step. There is often preprocessing, data checking, analysis, analysis checks, visualization checks, etc., etc., before the final result, which is usually either a data product or set of decisions and their supports.