Top Five Tricks for Coding in Pandas — with Matt Harrison

From the SDS 557: Effective Pandas — with Matt Harrison

[alwayz]

I'm a R tidyverse user who switched to Python. I've been chaining most of my pandas expressions because that comes natural to me (as a tidyverse user). I chain it without the parenthesis, but rather use backslash to go to new line. It has pros and cons. The main pro is I can select partial code and run it in VSCode interactive similar to that in R. It's still not a replacement for R but it's close. It looks like this:

df\
.assign(value2 = lambda df_: df_['value']**2)\
.groupby('type')\
.agg(value2_sum = ('value2', 'sum'))\
.reset_index()

I can select the code upto and including the first assign, run that first, check the results, and then run the rest. In this case, it's a trivial code. But in a complex chain, this can be very helpful.

[mabenba]

Amazing clip. I love the idea of chainning that Matt presented. I come from R as well and it truly feels like piping.

[Kinnoshachi]

Amazing! My two favorite internet teachers !!

[CaribouDataScience]

The number on trick.. switch to R and tidyverse :)

[peeintea]

Intro is terrible, way too loud and the sound is annoying. Please change it.
But the conversation is great.

[jbj]

Chaining isn’t that deep. So much exaggerations going on in the first half of the video. I chain when it make sense and don’t when it doesn’t matter which is basically all the time for all of data science.

[richardbennett]

Good stuff. The guest says "parenthese" for the singular form of parentheses when in fact, the correct form is parenthesis. He might have missed that point in elementary school.