Using plotmath in variable names

[sinclairjesse]

Is there a way of changing the variables names such that plotmath expressions could be used? I've tried adding a labels argument with a vector of new labels thinking that it may be passed through to text, but that didn't seem to work.

[vsimko]

Could you elaborate more on this ?
Or provide some example code-snippets, how it should look like.

[vsimko]

In the meantime, I found this stackoverflow question: http://stackoverflow.com/questions/35965433/r-corrplot-change-data-labels/36190501

[sinclairjesse]

Ideally, I would be able to use plotmath expressions (e.g., use of subscripts or superscripts) as a variable name, similar to using an expression in the axis labels in base graphics. For example:

plot(y ~ x, xlab = expression("Test"["A"]), ylab = expression("Test"^"B")).

Based on your answer in the stackoverflow question, colnames/rownames are used in the plotting. Something like:

mat <- colnames(expression("Test"["A"]), expression("Test"^"B"))

won't work.

[vsimko]

Now I understand.
I tried to implement plotmath for row/column labels but then I realized that I need some input from @taiyun - which approach is better ?

The problem is that colnames() and rownames() functions always convert their input to character even if the variable names are encoded using expression.

Approach 1: colnames always parsed as expressions

• Pros: simple to use
• Cons: variable names such as "alpha" will always be rendered as greek symbol alpha

M <- cor(mtcars)[1:5,1:5]
colnames(M) <- c("alpha", "beta", "alpha+beta", "a[0]", "a[beta]")
corrplot(M)

Approach 2: new parameters xlab & ylab for the corrplot() function

• Pros: control over plotmath expressions
• Cons: two ways for variable names specification. (1) dimnames from the original matrix, (2) xlab, ylab parameters which overrides colnames/rownames values.

corrplot(M, xlab = c("alpha", "beta",
                     expression(alpha+beta),
                     expression(a[0]),
                     expression(a[beta])))

Approach 3: similar to Approach 1 but using some prefix for expressions

Pros: simple to use
Cons: introducing special notation EXPR:...

M <- cor(mtcars)[1:5,1:5]
colnames(M) <- c("alpha", "beta", "EXPR:alpha+beta", "EXPR:a[0]", "EXPR:a[beta]")
corrplot(M)

[taiyun]

Thanks.
I prefer approach 3.

[vsimko]

OK, if I use Approach 3, what would be the best prefix:

• EXPR:
• MATH:
• E:
• or something else ?

[taiyun]

just:? eg ":beta"

[vsimko]

Other options would be:

• $ used in latex as $math expression$
• = used in spreadsheets for formulas

[taiyun]

I think anyone in $, : or = is OK.

[sinclairjesse]

Good solution! Thanks

[jonnez]

How do I now create a column or row name that contains just a $?

I currently get the following error (in version 0.84):

Error in ifelse(grepl("^[:=$]", s), parse(text = substring(s, 2)), s) :
  unimplemented type 'expression' in 'length<-'

[vsimko]

just use " $", e.g.

M <- cor(mtcars)
rownames(M)[1] <- " $"
corrplot(M)

[jonnez]

awesome! thanks

…

On Thu, Oct 5, 2017 at 9:49 PM, Viliam Simko ***@***.***> wrote: Closed #20 <#20>. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#20 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AGpo9hdGUV82mccuK3AD4ERY-Ho57b5Lks5spTKygaJpZM4HJzyk> .

[gezza1232]

Hey guys! I need to write MIP-1B as a column label, where B is a beta. I can't seem to be able to merge the MIP-1 with the :beta in order to get this to look right. Do you have any ideas on how to do that?

[vsimko]

what about this:

library(corrplot)
M <- cor(mtcars)
rownames(M)[1] <- ":MIP-1*beta"
corrplot(M)

[gezza1232]

It worked! Thanks

[Raed-Hamed]

Hi,

I believe your solution doesn't work whenever you want to have more than one subscript associated with a variable. Is that true and is there a way to fix this?

Thanks

e.g.

M <- cor(mtcars)[1:5,1:5]
colnames(M) <- c("alpha", "beta", ":alpha+beta", ":a[1,2]", ":a[beta]")
corrplot::corrplot(M)

Also the following doesn't work:
colnames(M) <- c("alpha", "beta", ":alpha+beta", ":a"["1,2"], ":a[beta]")