A Data Science Application (DSA) follows standardized steps, from initial project proposal/requirements to final project deliverables. This is the normal flow for any Data Processing application. Specific to DSA are the steps in between.
Caretakers of an end user application initiate a project in order to produce a desired result. This can be as simple as a Report, to the creation of a new system. For a DSA, this will most likely be the creation of datasets (files, database tables/SQL results) for use in Analysis/Reports as well as feeds to external processes.
Business Analysts utilize their knowledge, as well as attained knowledge from Subject Matter Experts (SME), in order to produce a Requirements document. This is a non-technical description of what is being requested for a desired result, as well as information/data sources needed for the result(s). These data sources can be anything from existing reports/online screens, to named data entities from a data dictionary/database catalog.