Antti Rask
youcanbeapirate.bsky.social
Antti Rask
@youcanbeapirate.bsky.social
Data Visualization | (Power) BI | R: RandomWalker package (co-author), TuneTeller Shiny App (author), and the author of the upcoming book ggplot2 extended l Helsinki Data Week (founder)
Sure, but I don't see this as an either/or type of situation. More as a both/and kind.

What using the pipe (in addition to the assignment arrow) gives me is the ability to decide what to assign and when.

Without the pipe, every step needs to be assigned and I've seen that lead to hairy situations.
December 15, 2024 at 9:12 AM
Anyway, here's the only change you need to make to get that. Not a big change, after all:

Before:
...
select(color, hex)
...

After:
...
pull(hex)
...
December 15, 2024 at 12:39 AM
Why return a tibble? Personal preference, I guess.

If I need to do something else, I’ll modify the code accordingly.

I’m more interested in showing what is possible instead of trying to nail the one true/right/correct/best way of doing a thing.

There are enough of those kinds of people around…
December 14, 2024 at 11:31 PM
And the alternative? I’m genuinely curious. Because for me the hardest code to read is when there are multiple functions inside one another.

And that’s the world (Excel) I came from when I started to learn R in 2020 (I know, I’m still a relative newcomer). But still I prefer the pipe.
December 14, 2024 at 11:02 PM
Not hating on data.table, but I am loving tidyverse.

And honestly, to me, clear and logical syntax is everything. Also, I'm not saying that tidyverse has gotten it right every time. But most of the times, yes.

ps. Ceci n'est pas une pipe... but yeah, pipes are part of the clear and logical syntax.
December 14, 2024 at 10:38 PM
Reposted by Antti Rask
You are not "too old," you don't "just not get it." Things have stopped being made intuitively by default. You are the victim, you are being gaslit into believing you are deficient when using software made by companies that have failed you.
November 29, 2024 at 7:36 PM
Actually, I think I changed my mind about DataExplorer. The part of it that deals with EDA is among the more useful visual summarization tools out there, it seems. Plus I'm experiencing some annoying problems with gt + gtExtras (especially the gt_plt_summary() function)...
November 28, 2024 at 5:11 PM
ggquickeda: this is the EDA app I was already going to write about
trelliscopejs: will be featured in the Interactive Plots section
ExPanDaR: this was new to me, but looks promising! Might find its way to the Automated EDA app chapter...
November 27, 2024 at 11:00 PM
GGally: this is one of the packages I will showcase in the Visualizing data subchapter
ggforce: this will be featured heavily in the book, but not in the EDA section
corrplot: not a ggplot2 extension, but I'm featuring ggcorrplot instead
November 27, 2024 at 11:00 PM
naniar: as mentioned, it was mentioned in the latest published chapter of the book
explorer: another new one to me, looks more like a general EDA tool than something specifically for visualizations
DataExplorer: similar to tidyplot, DataExplorer seems more of a Helper than an EDA tool to me
November 27, 2024 at 10:59 PM
@ercbk.bsky.social, here are my thoughts about the packages you mentioned so far:
tinyplot: not a ggplot2 extension as you mentioned, looks cool otherwise
tidyplot: this was new to me, but I would put it under the Shortcuts > Helpers section of the book instead of EDA
November 27, 2024 at 10:57 PM
@ercbk.bsky.social, I just finished writing the first version of 2.2.1 Missing values. Not surprisingly naniar is the main package I write about: ggplot2-extended-book.com/data#sec-mis...
ggplot2 extended - 2  Data
ggplot2-extended-book.com
November 27, 2024 at 9:37 PM
Thank you, @ercbk.bsky.social! Most of these were already on my radar, but I’ll check the complete list tomorrow to see if there were some I’ve missed previously. Much appreciated!
November 24, 2024 at 12:29 AM