Seaborn PairGrid | Compare PairGrid vs pairplot and use seaborn PairGrid map and map_diag

The seaborn PairGrid allows you to plot relationship plots as well as marginal distribution and is more flexible than the pairplot. In this video we compare the PairGrid vs pairplot, beginning by answering, "What is the PairGrid?" I then show you how to set up your PairGrid and then use pairgrid map to map various plots of your data onto the pairgrid. We then discuss the anatomy of the PairGrid and how you can use map_diag to map plots to the main diagonal, map_offdiag to map plots to the off-diagonal, and map_lower and map_upper to put plots on the lower and upper triangles of the pairgrid. Next, you'll see how to use the seaborn pairgrid hue to use color to represent a categorical feature in your dataset, and I also show you how to use pairgrid vars to specify only a subset of your numerical columns. Finally, we discuss pairgrid styling including how to change the pairgrid size with the arguments height and aspect. I also show you how to add a pairgrid legend.

0:00 Introduction
0:12 What is the PairGrid?
0:48 PairGrid Setup
2:28 PairGrid map()
4:32 PairGrid map_diag(), map_lower(), etc.
8:13 PairGrid hue and vars
10:49 PairGrid Styling
13:27 PairGrid legend
14:55 BLOOPERS!

// Github CODE //

// Related VIDEOS //

#seaborn #dataviz

[benkouti]

this is just best of the best, great job Kim. can't say this by words, but this video was very helpfull, so thanks very very much. best wishes Kim!!

[vitorribeirosa]

Simply fantastic. Thank you very much.
your videos are the ace of tutorials.
Cheers!!!

[jagan]

You have a unique style of explaining the Data Viz. diagrams. You have a great persona to appeal to a huge crowd with these wonderful nuggets. Thank you! Keep posting more. The bloopers were fun

[ssaa]

I am not surprised no one disliked the video. Thank you very much

[abhinavsharma]

Very Nice and informative video Kimberly

[chyldstudios]

Very good explanation of the PairGrid!

[rskura]

Great video Kim! Well done, as always. It's really nice to watch these and think about how I might use it with my own work. Please keep producing these high quality videos.

[mahmoudhamza]

I miss ur videos..

I have been waiting so long for the weekly videos. any news when they are coming back?

You have taught me lots and lots and things

I do wish you are doing well.

[guilhem]

Thank for the demo. Don't hesitate to make other videos like that with other sns methods. Python libraries are a jungle!

[sakthis]

Very Good Explanation... Thank you for this great tutorial for beginners :)

[youcefyahiaoui]

Hi Kimberly, As always, it will be difficult for me to articulate how outstanding your videos are. Suffice to say a sincere thank you for the super valuable work you do. I may have asked this question before but I would love to see a video about the structure of the data these grid and "automated" methods like PairGrid and their map functions that are outstanding and very flexible. But I am running into issues when my data are simply columns of variables (I don't have any categorical variable). Say I want to specify which columns to use with x_vars and which I can use for y_vars. Then how can I use map methods to do what you showed and how, for instance, would I put texts to individual formed subplots? If I used PairGrid on my whole dataframe, then it will form a grid corresponding to all the dataframe. it will decide to look at cross correlation for all the columns. If I don't want to do that, I want to select columns to pairplot with only one other column (with wrapping the grid), how would I do that?

[BiologyIsHot]

I'm super lost about what gets passed by map_lower() and map_upper(). For instance, if I try to wrap a modified version of sns.lmplot() I get an error that the required "data" argument is empty

[piusnnwachukwu]

Thank you for such a nice video, just that I still have difficulty interpreting the histplot. What exactly it is representing

[ramdevprajapati]

i have watch your all the videos but i don't understand why are you using ;