pandas best practices (3/10): Comparing groups

This video covers the following topics: filtering a DataFrame, value counts, normalization, groupby.

NEW TO PANDAS? Watch my introductory series (30+ videos):

DOWNLOAD the dataset and notebook:

SUBSCRIBE to learn data science with Python:

JOIN the "Data School Insiders" community and receive exclusive rewards:

LET'S CONNECT!

[kevinz]

These videos are absolutely fantastic. Everything is explained in a way that it extendable to countless other problems. Hate to be that guy ...but I have to jump on the bandwagon and mention that I hope you keep more videos coming! Even the way the video series is titled has helped me organize in my head how I think about approaching a problem using pandas. Thank you so much for the time you put into this.

[brentskoumal]

super helpful Kevin, always love watching your tutorials, keep them coming!

[Kristina_Tsoy]

Thank you Kevin! All your videos are super helpful👍

[trungthanh]

Thank you. Good lecture, best delivery

[WaylonWalker]

Love the unstack at the end! That is one of my favorite commands.

[joseluisvallespardo]

Nice videos!!
Is there any way to show at the same time the normalize and the absolute result?

[crigar]

when i put this: ri[ri.violation == 'speeding'].driver_gender.value_counts(), out is:Series([], Name: driver_gender, dtype: int64)

[soobrazil]

Another way to do the same, hope it is pythonicly correct:
pd.crosstab(ri.driver_gender, ri.violation).Speeding

[alexlfo]

I was thinking to ri[(ri.driver_gender=='M') & (ri.violation=='Speeding')]. But is that passaible to become a number data instead of DateFrame by adding some method ?

[akshayakn]

and its output is empty column..what could be the problem

[sarikadatta]

Grouping by gender will do the trick too

[lenahudson]

If we want to drop columns having more then 30% of NA (or more then 100 NA)