3

I am a student getting started with Tableau for the first time. My proficiency with ggplot2 is intermediate. I can create custom versions of the most popular kinds of charts in ggplot2 but nothing too fancy (and not very time-efficiently).

I am interested in creating more polished visuals of my data in Tableau. My ggplot2 ones aren't bad, but they are not what I would call aesthetically pleasing either.

What I would like to know is under what situations I should stick with ggplot2. The only advantages that come to mind with my current (albeit limited) understanding of Tableau is to keep my work reproducible and open source. ggplot2 is also probably more customizable, but I am continually impressed by Tableau's offerings as I learn to use it.

This question is not intended to be a flippant jab at ggplot2. As I did not hear any talk about Tableau in the academic circles I was in, I am wondering what obvious drawbacks to Tableau may exist that would explain its unpopularity there.

SKOR2
  • 33
  • 1
  • 3

2 Answers2

5

There is one big economic difference between the two: ggplot2 is an open source package for an open source programming language. On the contrary Tableau is a proprietary software. That might be a dependency that you might not want to risk, e.g. you do everything in Tableau but then the license gets more expensive and your organization does not want to spend money for it anymore. Academics would often stress this reason, another reason of low adoption there is that Tableau has always been the BI tool and R the "research" tool.

Apart from this, they are simply different tools (in the same domain of data viz) and as such will be suited for different problems. This can be subject to personal preference, but I would say that there is a few basic considerations:

  • Speed: how fast can you get what you need?

    • Use the tool that gets you results. Tableau might often be faster.

  • Reproducibility & automation:

    • If you know you might want to plot this chart many times in future in a consistent manner or embed your plots in some sort of notebook like R Markdown or dashboard, then R is most likely the better choice. You write the code once and produce thousands of plots within seconds if you need to.

  • Type of project:

    • If you are starting with messy data, or you are doing some modeling work, it seems as a good strategy for me to just stick to R, I see it as more efficient to have everything in one place.
    • If you need to get awesome visualizations from clean data, Tableau will excel
    • Based on the type of visualizations and output you want, one or the other might be better, ggplot2 is extremely flexible, Tableau has a lot of amazing functionality which is ready to go.
František Kaláb
  • 76
  • 2
1

This doesn't answer your Tableau questions, but I can say a few things about ggplot2. ggplot2's has reasonable defaults but nearly everything about the plots can be changed to make the results more clean aesthetically. One package I use all the time is called 'cowplot': https://cran.r-project.org/web/packages/cowplot/vignettes/introduction.html

ggplot2 also has built-in themes (http://ggplot2.tidyverse.org/reference/ggtheme.html) and there's a package for community-develeoped themes (https://cran.r-project.org/web/packages/ggthemes/vignettes/ggthemes.html).

tom
  • 2,228
  • 11
  • 13
  • I know how to change the themes, but I'll be sure to check out cowplot. Thanks for your response – SKOR2 Mar 27 '18 at 18:13