Orchestrating Large and Small Projects With Apache Airflow

Have you worked on a project that needed an orchestration tool? How do you define the workflow of an entire data pipeline or a messaging system with Python? This week on the show, Calvin Hendryx-Parker is back to talk about using Apache Airflow and orchestrating Python projects.

Calvin is the co-founder and CTO of Six Feet Up and a Python Web Conference co-organizer. He’s recently been working on a massive project that requires thousands of jobs involving transferring and transforming data. Through his research into orchestration systems, he found Apache Airflow.

Airflow is an open-source tool to define, schedule, and monitor workflows. The platform is pure Python and integrates with a wide variety of services. We discuss how workflows are defined by creating directed acyclic graphs (DAG).

Calvin talks about how a recent project outgrew the system and how his team built a clever solution using Python. We also discuss the upcoming Python Web Conference and what virtual attendees can expect.

Topics:

• 00:00:00 – Introduction
• 00:02:24 – Describing the large data pipeline
• 00:04:38 – What format was the data in?
• 00:06:04 – Was the format of the data changed for storage?
• 00:09:34 – Data engineering and describing sources and targets
• 00:11:29 – Apache Airflow orchestration and hitting limitations
• 00:18:12 – Sponsor: CData Software
• 00:18:54 – DAG: Directed acyclic graphs
• 00:22:29 – Streaming data and other tool choices
• 00:25:38 – Overcoming DAG Factory limitations
• 00:31:49 – Another industry example for Airflow
• 00:34:24 – Finding solutions as a consultancy
• 00:35:12 – Is there a minimum-size project for Airflow?
• 00:37:37 – Django under the hood
• 00:38:31 – Video Course Spotlight
• 00:39:58 – The Python Web Conference 2023
• 00:44:24 – Do you have any upcoming conference talks?
• 00:45:53 – How can people follow your work online?
• 00:46:52 – IndyPy talk by Mariatta Wijaya
• 00:48:01 – What are you excited about in the world of Python?
• 00:51:45 – What do you want to learn next?
• 00:53:22 – Thanks and goodbye

Show Links:

• Apache Airflow - Documentation
• Too Big for DAG Factories? — Six Feet Up
• Directed acyclic graph - Wikipedia
• DAGs — Airflow Documentation
• Dynamically generating DAGs in Airflow - Astronomer Documentation
• Data Lakehouse Architecture and AI Company - Databricks
• Episode #10: Python Job Hunting in a Pandemic – The Real Python Podcast
• Episode #124: Exploring Recursion in Python With Al Sweigart – The Real Python Podcast
• The Recursive Book of Recursion
• Episode #61: Scaling Data Science and Machine Learning Infrastructure Like Netflix – The Real Python Podcast
• IndyPy — Indiana Python User Group
• Contributing to Python - Mariatta Wijaya - Python Core Developer - YouTube
• Home Assistant
• Arturia - MicroFreak
• Arturia - Pigments
• CalvinHP (@calvinhp@fosstodon.org) - Fosstodon
• calvinhp - Twitter
• Six Feet Up - Blog
• Python Web Conference 2023

Level up your Python skills with our expert-led courses:

• Data Cleaning With pandas and NumPy
• Python Basics: Object-Oriented Programming
• Intro to Object-Oriented Programming (OOP) in Python

Support the podcast & join our community of Pythonistas