Data Pipeline Frameworks: The Dream and the Reality

Published on

One of the core problems in data engineering is defining and orchestrating scheduled ETL pipelines. While aspects of the problem are general, the dream is to choose and use a framework that does “everything but write your query.”

In reality, frameworks are useful but do less than they promise.  Your team still has to do the rest of the work required to fit the framework to your business, existing code, and ops practices.

Hear more about what I learned from these tradeoffs in my first phase of working with open source pipeline framework Airflow in my talk for Data Council NYC 2018.