- They tried too hard to do a Tuft* - but it's actually way too flashy.
- Basically the issue is that it's not generic enough. Whenever you see a ggplot2 plot it's always screaming at you "I WAS MADE IN ggplot2!!" Just KISS...
- The grey background is completely unusable if you want to print out your charts!
- The default pastel colors (ie. PowerPoint 2012) are always a disaster for readability. In meetings people constantly can't tell them apart and they are definitely not colorblind friendly.
It' a shame b/c each script I write ends up having to have an obnoxiously long theme declaration to make it look "normal"
* I think most people miss the Tuft's point. He presents a way about thinking about data presentation. Instead people just look at it and think "oooo! That looks pretty. Let me copy what he did". The guy has his own personal style - it's definitely nice, but the point isn't to copy it.
EDIT: For those thinking about learning ggplot2.. maybe wait off for a bit. It seems like it'll be deprecated soon and replaces with ggvis.
I would recommend using theme_bw(), which helps solve some of the problem with the gray background.
EDIT: This is ggvis: http://blog.rstudio.org/2014/06/23/introducing-ggvis/
This is news to me, so I'll look into it to see how it differs from ggplot2.
EDIT 2: It seems more like that the difference is that ggvis is more for interactive charts, but then it requires a dependency on Shiny, which is not optimal for blog posts.
I'm definitely not an R guru or "in" on the latest news, but it seems that Hadley Wickham (who probably single handedly is the reason R is still relevant) now works for the RStudio guys and he's reworking his tools. plyr is now dplyr and gpplot2 is now ggvis. And there also another tool called tinyr. My understanding is that they're still in development, but they'll ultimately provide an "integrated" ecosystem for processing data.frames (with hooks into the RStudio IDE)
He talks about it at the beginning of this video https://www.youtube.com/watch?v=wki0BqlztCo
https://www.coursera.org/specialization/jhudatascience/1?utm...
http://stackoverflow.com/questions/12675147/how-can-we-make-...
You can customize ggplot2 for even more beautiful charts, such as the ones used in my analysis of HN comments: http://minimaxir.com/2014/10/hn-comments-about-comments/
Note that reproducing this analysis for the HN front page is harder since there is a lot that happens behind the scenes that cannot be generalized. Here's an attempt at performing such an analysis (note dang's comments): https://news.ycombinator.com/item?id=8533757