I am submitting this question to HN to perhaps collect more diverse perspectives than I could get from language models and search engines, given the audience here. I have tried to lean into using Visidata[2] a bit, but it has been a steep learning curve so far, and it hasn't exactly been easy to take visidata transformations and apply them to our python codebase.
1: https://devblogs.microsoft.com/python/data-wrangler-release/
2: https://www.visidata.org/