Hi @arfon

You have helped me with this once before (jennybc/googlesheets#98) and I see you in various threads in https://github.com/github/linguist.

Can you figure out how to make this recognized as an R repository? I need to ignore lots of README.md and README-fodder.md files in subdirectories.

I've tried everything I can think of with my .gitattributes file, with no success.

Thanks!

I need to mine my own ggplot2 tutorial, for example:
https://github.com/jennybc/ggplot2-tutorial

This R Chart Chooser from 2012 is very good, but it doesn't help you get the plot and code in front of your eyeballs at same time:
http://www.yaksis.com/posts/r-chart-chooser.html

This ggplot2 cheat sheet is closer to what we've done here but focuses more on incremental changes to details and is formatted like a blog post vs. a clickable quilt:

http://zevross.com/blog/2014/08/04/beautiful-plotting-in-r-a-ggplot2-cheatsheet-3/

Hello, really nice project here!

I'm wondering whether the developers would be open to giving permission for this to be incorporated in a directory for other languages and packages? There's no LICENSE file, apologies if I've missed it somewhere - "We are very interested in extending this catalog" on the about page sounded promising so I thought I'd ask.

Thinking R (ggplot2), MATLAB, a select few of the many Python libraries (matplotlib / seaborn / ggplot / bokeh / vincent), Julia (Gadfly) etc.

Could be an interesting [potentially rather big] challenge but shouldn't be too difficult for any part individually, and would make it easier for people to move out of their comfort zones. The R catalog could make for a good set to replicate in these languages.

Example online, for a single bar chart: badhessian.org/2014/07/six-of-one-plot-half-dozen-of-the-other (R ggplot2, R base, Python matplotlib, Python seaborn, Julia Gadfly, and Plot.ly)

Hi @jennybc
The link to launch R Graph Catalog doesn't appear to be working anymore. I believe I remembered you showing this during STAT 545 and I was hoping to show it to a grad module I'm teaching at UBC in September (BIOL 548L) so they can quickly get a glimpse of a whole bunch of different charts + code. Am I just missing something?
Thanks!
Melissa

This only happens in the stats Shiny server, but not locally

Following the ubc stats link leads to the error

Error in shinyAppDir(x) : 
  No Shiny application exists at the path "/nfs/37zfs1-shiny-server/shiny-server/r-graph-catalog"
Calls: runApp ... as.shiny.appobj -> as.shiny.appobj.character -> shinyAppDir
Execution halted

I guess UBC people aren't hosting this anymore?

From the readme, http://shiny.stat.ubc.ca/r-graph-catalog/ is reporting that the site can't be reached. isitdownrightnow.com confirms that the site is down.

This can be done by adding a URL query param eg. "http://shiny.stat.ubc.ca/r-graph-catalog/?fig=figure3" and inside the server code, if the URL param is detected then simulate a click to go to that figure

Hi Jenny. Good day!

I have been trying to reach the R Graph Catalog website for the past couple of days and I get the following message:

This site can’t be reached

Any assistance would be greatly appreciated.

Thanks!

I just re-made all the figures (5744c62) and the version of, e.g., ggplot2, among other things, should get automatically recorded somewhere.

Context: I'm using make init to add a few figures. The first one is taking a while to develop but the second one was easy. When I do make all, the code for the first figure is run, but throws an error. Of course I can fix the immediate problem by commenting out the code, but I suspect other downstream problems will follow.

Makes me realize we need a way to shut off processing for an entire figure (directory), while leaving what's there intact.

This seems related to the not doing tag, which seems to come into play quite late (perhaps only when preparing the app?). Better (= earlier) handling of that might be the solution.

If you try to click on a figure that's already the last viewed figure, nothing happens (you don't get automatically moved to the figure tab)

So far, all the reshaping tasks I've encountered (I'm in chapter 4) look well within the capability of tidyr. Let's switch over to that instead of explicitly using reshape2. It's a friendlier package to present to people. (Of course, under the hood, reshape2 is still being used.)

As I go through the figures, it's becoming clear that we need a single definitive source for several datasets.

Example: the museum visit length data from Fig 4.10 + others.

It needs to exist in one single location, probably in short and tall form.

This underscores how nice it would be to create a companion nbr or cmeg data package, even if it's data we have simulated that behaves like the original data we see in these figures.

Then, instead of copying the data around to multiple figure directories, we could just load it with a call to library(cmeg).

Side bonus: we could set factor levels sensibly.

Received by @jennybc via email 2015-02-03:

I wrote this email just to say "Thank all of you" and remind you that the page may have a display bug under Firefox: Catalog is too wide to see the third column. It seems that some graphs like Fig 2.2 cannot zoom properly.

This only happens in the stats Shiny server, but not locally