I do use Jupyter and the boilerplate code above is what I usually deploy in Jupyter. Spark syntax is still just a touch more complicated than necessary for my

I do use Jupyter and the boilerplate code above is what I usually deploy in Jupyter. Spark syntax is still just a touch more complicated than necessary for my needs because it's solving the much more general distributed computing problem. (not to mention the impedance mismatch -- Spark is very much a Scala native app, and it feels that way to a Python user). DuckDB on the other hand seems to provide a quick way to interact with local Parquet files in Python, which I appreciate.