How to recreate an Excel line graph with a table with ggplot

[Timothy Lau]

How do should I go about recreating this graph in ggplot?

[Brandon Hurr]

Tim

You're gonna have a difficult time getting the exact same output but you can get close with the following.

Get the data into an R dataframe.
Create the top plot with geom_point, geom_line and geom_text.
Then create a table with grid and put the two together with the gridExtra package.

I remember other people doing something similar so search stackoverflow and the archives here for help.

If you're stuck, supply a dataset in dput() format and your attempt at getting started.

B

On Wed, Apr 29, 2015 at 18:08 Timothy Lau <timoth...@gmail.com> wrote:

How do should I go about recreating this graph in ggplot?

--
--
You received this message because you are subscribed to the ggplot2 mailing list.
Please provide a reproducible example: https://github.com/hadley/devtools/wiki/Reproducibility
 
To post: email ggp...@googlegroups.com
To unsubscribe: email ggplot2+u...@googlegroups.com
More options: http://groups.google.com/group/ggplot2

---
You received this message because you are subscribed to the Google Groups "ggplot2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ggplot2+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[John Kane]

Well ideally you would not. That plot (mutant ?) shows why one should not even think about using spreadsheet graphics. There is way too much 'chart junk',. aka 'crap', there. Google “Tufte chart junk” for a discussion

Generally one need a plot of the data or a table but not both and labelling the numeric data points in the actual plot is overkill, that is why there is a y-axis.

If you feel that you really need a table of the data as well as the graph (or vis versa for that matter see below.

##=========================================================
char1  <-   melt(chartjunk, id.vars = c("xx"))
g1 <-  ggplot(char1,  aes(xx, value, colour = variable )) +
             geom_line() + geom_g1  point()
g2 <- tableGrob(chartjunk)

grid.arrange(g1, g2, ncol=2)

##=========================================================

If you really need numbers on the graph itself us
g1 <-  ggplot(char1,  aes(xx, value, colour = variable )) +
             geom_line() + geom_g1  point()  +
            geom_text( aes(label = value))

[John Kane]

Blast, only pasted half the code. Here it is in a form that might even work. Apologies.

library(RGraphics)
library(gridExtra)
library(ggplot2)

chartjunk  <-  data.frame(xx = 16:35,
                  act = c( 1,5,8,42,192,171,143,114,102,133,100,94,69,69,51,34,18,12,7,1),
                                dg = c(0,1,1,3,12,20,17,18,21,38,29,34,30,28,22,19,11,9,6,1))

char1  <-   melt(chartjunk, id.vars = c("xx"))
g1 <-  ggplot(char1,  aes(xx, value, colour = variable )) +
             geom_line() + geom_g1  point()
g2 <- tableGrob(chartjunk)

grid.arrange(g1, g2, ncol=2)

On Wednesday, April 29, 2015 at 9:08:19 PM UTC-4, Timothy Lau wrote:

[John Kane]

At this rate I'll never get it right.1 
Add  library(plyr) to the code
Apologies again.

jrk

On Wednesday, April 29, 2015 at 9:08:19 PM UTC-4, Timothy Lau wrote:

[jwd]

On Wed, 29 Apr 2015 18:08:19 -0700 (PDT)
Timothy Lau <timoth...@gmail.com> wrote:

>
>
> How do should I go about recreating this graph in ggplot?

> <https://lh3.googleusercontent.com/-s_Aj4i1PyZ4/VUF_3kUJRzI/AAAAAAAB7pU/A0MvwGlUyJY/s1600/great%2Bgraph.PNG>
>

Ideally, you shouldn't. The chart is misleading, and the table is
likewise for any reader that doesn't look closely at the numbers.
There is a strong positive correlation between ACT score and the number
of degrees awarded as a proportion of those tested and receiving that
score that is not apparent on the chart.

JDougherty