Alignment and scaling issue

[atkinsjeff]

Follow-up from Twitter conversation...

Trying to make multiple waffle charts, in this case three of them, of raw abundance count, for comparison of abundance against the individual site. Example below is for brook trout abundance in three streams:

part one making the individual plots:

> pain.adult.1997 <- c( `YOY (406)`=406, `Adult (24)`=24)
p.pain.waffle.1997 <- waffle(pain.adult.1997/2, rows=7, size=0.5, 
       colors=c("#c7d4b6", "#a3aabd"), 
       title="Paine Run Brook Trout Abundance (1997)", 
       xlab="1 square = 2 fish")

> pine.adult.1997 <- c( `YOY (221)`=221, `Adult (143)`=143)
p.pine.waffle.1997 <- waffle(pine.adult.1997/2, rows=7, size=0.5, 
                             colors=c("#c7d4b6", "#a3aabd"), 
                             title="Piney River Brook Trout Abundance (1997)", 
                             xlab="1 square = 2 fish")

> stan.adult.1997 <- c( `YOY (270)`=270, `Adult (197)`=197)
p.stan.waffle.1997 <- waffle(stan.adult.1997/2, rows=7, size=0.5, 
                             colors=c("#c7d4b6", "#a3aabd"), 
                             title="Staunton River Trout Abundance (1997)", 
                             xlab="1 square = 2 fish")

I tried a back door method that I have found works well with typical plot elements using the gridExtra and gtable packages to arrange plots. This code is pretty rough. This actually fixes the problem of scaling, but then centers everything.

attempting to gride the waffles

require(gridExtra)
require(gtable)

pain panel graph

gA7 <- ggplot_gtable(ggplot_build(p.pain.waffle.1997))
gB7 <- ggplot_gtable(ggplot_build(p.pine.waffle.1997))
gC7 <- ggplot_gtable(ggplot_build(p.stan.waffle.1997))

maxWidth = unit.pmax(gA7$widths[2:3],
gB7$widths[2:3],
gC7$widths[2:3])

Set the widths

gA7$widths[2:3] <- maxWidth
gB7$widths[2:3] <- maxWidth
gC7$widths[2:3] <- maxWidth

Arrange the four charts

grid.arrange(gA7, gB7, gC7, nrow=3)

This may be a rather crude way to do it if one wants to make multiple graphs (belaying my abject novice status here). But interested in how to best go about this.

[hrbrmstr]

Perfect examples. Very much appreciated. I pushed (literally minutes ago) a version with many nulled margins. Give that one a go if you can (it won't fix all the issues). I ran that updated version with your code and did a grid.arrange of the output.

The charts aren't left-aligned, but the blocks are the same size at least.

[atkinsjeff]

Oh that's awesome. Thanks. I am going to play around with the left-align
and see if I can get that to work. Some threads in stackoverflow lead to
dead ends.

--jeff

On Thu, Mar 19, 2015 at 11:52 AM, boB Rudis [email protected]
wrote:

Perfect examples. Very much appreciated. I pushed (literally minutes ago)
a version with many nulled margins. Give that one a go if you can (it won't
fix all the issues). I ran that updated version with your code and did a
grid.arrange of the output.

[image: image]
https://cloud.githubusercontent.com/assets/509878/6734404/59e6b5e0-ce2e-11e4-92da-e6cf488fcf2e.png

The charts aren't left-aligned, but the blocks are the same size at least.

—
Reply to this email directly or view it on GitHub
#4 (comment).

Jeff Atkins
Ph.D. Candidate
Department of Environmental Sciences
University of Virginia

[hrbrmstr]

I added a parameter pad to waffle to add pad columns of blocks. If you find the max width (I guess I shld write a small function for that, eh? ;-) in terms of # of blocks for the largest plot in a series, then pad the others to get to the same total width. Then use the new iron function to stitch them together. Like this:

pain.adult.1997 <- c( `YOY (406)`=406, `Adult (24)`=24)
A <- waffle(pain.adult.1997/2, rows=7, size=0.5, 
       colors=c("#c7d4b6", "#a3aabd"), 
       title="Paine Run Brook Trout Abundance (1997)", 
       xlab="1 square = 2 fish", pad=3)

pine.adult.1997 <- c( `YOY (221)`=221, `Adult (143)`=143)
B <- waffle(pine.adult.1997/2, rows=7, size=0.5, 
                             colors=c("#c7d4b6", "#a3aabd"), 
                             title="Piney River Brook Trout Abundance (1997)", 
                             xlab="1 square = 2 fish", pad=8)

stan.adult.1997 <- c( `YOY (270)`=270, `Adult (197)`=197)
C <- waffle(stan.adult.1997/2, rows=7, size=0.5, 
                             colors=c("#c7d4b6", "#a3aabd"), 
                             title="Staunton River Trout Abundance (1997)", 
                             xlab="1 square = 2 fish")


iron(A, B, C)

Let me know how that works for you.

[atkinsjeff]

This looks great. That the function is called Iron makes it even better!

--jeff

On Thu, Mar 19, 2015 at 7:32 PM, boB Rudis [email protected] wrote:

I added a parameter pad to waffle to add pad columns of blocks. If you
find the max width (I guess I shld write a small function for that, eh? ;-)
in terms of # of blocks for the largest plot in a series, then pad the
others to get to the same total width. Then use the new iron function to
stitch them together. Like this:

pain.adult.1997 <- c( YOY (406)=406, Adult (24)=24)
A <- waffle(pain.adult.1997/2, rows=7, size=0.5,
colors=c("#c7d4b6", "#a3aabd"),
title="Paine Run Brook Trout Abundance (1997)",
xlab="1 square = 2 fish", pad=3)

pine.adult.1997 <- c( YOY (221)=221, Adult (143)=143)
B <- waffle(pine.adult.1997/2, rows=7, size=0.5,
colors=c("#c7d4b6", "#a3aabd"),
title="Piney River Brook Trout Abundance (1997)",
xlab="1 square = 2 fish", pad=8)

stan.adult.1997 <- c( YOY (270)=270, Adult (197)=197)
C <- waffle(stan.adult.1997/2, rows=7, size=0.5,
colors=c("#c7d4b6", "#a3aabd"),
title="Staunton River Trout Abundance (1997)",
xlab="1 square = 2 fish")

iron(A, B, C)

[image: image]
https://cloud.githubusercontent.com/assets/509878/6743322/b3d85e60-ce6e-11e4-981f-5266c4e0c833.png

Let me know how that works for youl

—
Reply to this email directly or view it on GitHub
#4 (comment).

Jeff Atkins
Ph.D. Candidate
Department of Environmental Sciences
University of Virginia