Prefect.io

A library to put your data science into production.

Why Prefect and not Apache Airflow?

Honestly, I main chose it because the docs looked better and someone in my class made a contribution to project. Another more important reason is that small tasks seem easier with Prefect. I have a relative simple ETL job to run daily. I was able get the core operation running relatively quickly and the scheduler was also very easy to get started.

Tasks

The Prefect library uses the decorator @task a lot. The decorator turns your function into a class and gives it some prefect methods. So far I haven't had to access any of those.