Hacker News Reader

Anything unambiguous really, whatever works for the job. So unless the job is looking at very limited amount of data and drawing conclusions from it just using your brain, that means you need some kind of visalisation and processing tool. Like pandas or excel since we're talking tabular data. I.e. actually the whole premise of 'let's look at data in a plain text editor' is sort of wrong, imo. It's too limited and hence not the right tool for the job most of the times. You can't do anything but the simplest things with the data anyway in a plain text editor, so from the moment your data is more than a couple of rows in a table you're better off with another tool anyway.

JSONL (jsonlines.org)

http://jsonlines.org

pandas can directly write to parquet

parquet has a very narrow use case. it's great for internal operations but interoperability is limited unless both sides use one of the same handful of tools. csv is supported by literally every tool ever designed to work with data.