Why I don't use ggplot2 (simplystatistics.org)

Having trouble getting past the unreadable font and broken scrollbar in order to read this article about how difficult ggplot2 is to use.

Looks like the site is down.

I don't agree with the author on some of these points but its seems perfectly fine for what it is given that the title is "Why _I_ don't use ggplot2" not "Why _you_ shouldn't use ggplot2".

Site down, try the cache: https://webcache.googleusercontent.com/search?q=cache:http%3...

Using a very, very simplistic chart and dataset for comparing base graphics and ggplot2 is cheating a bit. Base graphics can become extremely unwieldy when dealing with nonunivariate data, especially if you want specific theming.

The ggplot2 code is not formatted one-function-per-line like the base code. It may be slightly more LoC for base graphics, but it's very clear what is happening with the ggplot2 and design, which is a far more important attribute than LoC.

The author makes an interesting point that ggplot2 defaults often look polished enough to make a novice find them acceptable when they're really not.

It's a beautifully polished and well-documented R package, but often I feel like I've wrestled a bear to get a figure just right. The effort though seems worth it for the results and I don't feel that way about the base graphics.

I suppose if you've already memorized the arcane syntax of base R's plotting functions, along with a giant laundry list of highly customized visualizations available via libraries, then perhaps ggplot2 really is slower for exploratory plotting. But for anyone coming to R fresh, the expressiveness and coherence of ggplot2's grammar absolutely blows away base R for fast exploratory graphics.

For working data scientists in industry, there is an extremely important middle ground between exploratory plots that only I will ever see, and publication quality plots. The most important plots I make are those that I'll share with a small audience within my company. These need to be clear and aesthetically pleasing, but not polished to an extreme degree. Again, ggplot2 excels in this regime.

actually making heatmaps is rather easy, checkout: http://docs.ggplot2.org/current/geom_tile.html

    ggplot(melt(matrix), aes(x = Var1, y = Var2, fill = value)) + geom_raster()

Nothing that basic archiving can't fix. He should just maintain a script containing his ggthemes of predilection for reuse.

I see ggplot2 is not so much as a replacement for base graphics but rather for lattice, and as such it is pretty amazing. Downsides of ggplot2 are the poor default theme and the fact that it's horribly slow. For bigger data sets it is often two orders of magnitude slower than base graphics, which can be prohibitive.

I use a bit less of the Hadleyverse than my colleagues like. But ggplot2 has a number of advantages over base graphics (even if you were to figure out base graphics). One being the plot is a structure or value (returned by plot construction) and not a bunch of side effects happening in a viewport. This orientation is much more compatible with functional programming (the author hints at this with the "compatible with piping" point).

I really like all plotting systems in R. First, I used base graphics for a few years—and loved it. You learn your way around par(), commit esoteric argument names to memory (oma, mar, mgp, mfrow, etc). It feels powerful — you're just drawing on a screen; its history traces to the original pen plotters. Second, I learned lattice. You can't help but fall in love with lattice after a year or two with creating panel plots in base graphics. The biggest learning curve with lattice is panel functions, but once you learn to throw a browser() in a panel function for stack variable introspection, you can do anything. Somewhere on a dusty bookshelf is a well-worn lattice book I splurged on while taking an R course at UCD.

I like this article, because I think for production graphics, the author has a point. If you're placing lines, points, and labels on a screen — you can create anything. You can draw polygons and arcs. It's like drawing with raw SVG. But I'd have a hard time thinking of an exploratory data analysis situations I wouldn't reach for ggplot2 first. Since it looks at dataframe column types (integers, factors, numerics), it automatically matches these two the appropriate type of color gradient. Coloring a scatter plot by a potential confounder is one additional argument to aes(), e.g. aes(x, y, color=other_col). More than once during EDA I've done this and seen some horrifying pattern in data that shouldn't be there. That's a powerful tool for one extra function argument — the cost of checking for a confounder with color (or shape) is essentially near zero.

I'd make the case that this is a more costly operation in base graphics, and is thus much less likely to be done. You may already have your plots in a for loop to create panels, then you have a few extra lines for adjusting margins and axes (rather than facet_wrap(~col)). It took a lot of code to set that up — there's already a lot of cruft when you just need to do a quick inspection. Then you need to create a vector of appropriate size of colors, and then map this to data. Sure it's easy-ish, but it takes at least double the time as color=some_col. In EDA visualization, I want every single barrier to checking a confounder to be as small as possible—which is what ggplot2 does.

That said, I really liked this article because I do agree that going from EDA visualization to production is a hassle. Just after reading this, I remade some production ggplots with base graphics and love the simple aesthetic — which to mirror in ggplot takes a lot of hassle.

What I really long for is a lower-level data to visualization mapping (like d3) in R. d3 is a pain to learn, but it's really the only data abstraction (even though it is a low-level abstraction) that is seemingly limitless in what it does and can do. I always hope for a general data-join grammar like d3's to be the norm, built on top of base plotting (analogously: svg elements), and then have abstractions like ggplot for tabular data built on top of that.

I don't just base graphics because making a simple legend is a PITA, when every other plotting library seems to handle it fairly easily.

Question to the author: have you tried a GUI like Deducer for creating ggplot2 graphs? If yes, what was your experience like? If no, why not?