aphalo / ggpmisc Goto Github PK

When I use stat_poly_eq with formula=y~I(1/x), the wrong equation is printed. In the case of the code below, it should be
y = 1/x,

but instead it gives:

y = -0.00033 + 1.01x

library(ggplot2)
set.seed(0)

tibble::tibble(x=1:100, y=1/x + rnorm(100, sd=0.01)) |>   
  ggplot(aes(x, y))+
  geom_point()+
  geom_smooth(formula=y~I(1/x), method="lm", se=F)+
  ggpmisc::stat_poly_eq(ggpmisc::use_label("eq"), formula=y~I(1/x), method="lm")

^{Created on 2023-06-29 with reprex v2.0.2}

stat_poly_line(), stat_poly_eq() and relatives should be able to gracefully handle NA and NULL returned by model fit methods as an indication of failure, so that in such cases no line or equation is added to the plot. This would be useful in user defined wrappers on model fit methods.

As done with 'ggpp', unit tests need to be added to 'ggpmisc' so as to test the code better (in actual practice) and to achieve formally high coverage.

Currrently we support lm() and rlm() and any other function that returns an object of class "lm". MASS::lqs() with method lts is a useful alternative that could be also supported rather easily, I think.

https://stackoverflow.com/questions/65695409/is-there-a-neat-approach-to-label-a-ggplot-plot-with-the-equation-and-other-stat

This would be usable with ggplot2::geom_quantile()

There are two problems happening when using stat_ma_line():

1. when I use the range.y and range.x = "interval", the ggplot code works fine when I use the variable "y = wind250_1". However when I use the variable "y= lee250_1" I get the error: Discrete value supplied to continuous scale.

2. if I change the range.y and range.x = "relative". The code runs fine for both y variables. However, when I exclude the lines " "scale_x_continuous(trans='log10')" and scale_y_continuous(trans='log10') and use "y=lee250_1", it gives me the error: Discrete value supplied to continuous scale.

The variable is numeric, I even used the as.numeric() in the code. The values per group could be a problem, as suggested in this post - https://stackoverflow.com/questions/75242355/error-discrete-value-supplied-to-continuous-scale-stat-ma-line. However it does not make sense for me considering that the code works with the variable "wind250_1" and not with "lee250_1".

ROS<- c(0.03, 0.03, 0.03, 0.03, 0.07, 0.07, 0.07, 0.07, 0.07, 0.1, 0.1, 0.1)
wind<- c(0.84, 1.77, 3.5, 6.44, 0.84, 1.77, 3.5, 6.44, 7.55, 0.84, 1.77, 3.5)
rey <- c(31500,66375,131250,241500,31500,66375,131250,241500,283125,31500,66375,131250)
wind250_1 <- c(69.4,69.4,1,1,31.08,37.07,1,1,1,22.8,19.45,1) 
lee250_1 <- c(79.84,125.56,93.34,94.42,33.78,49.6,38.95,40.9,39.32,24.2,32.95,27.46) 

df<- data.frame(ROS,wind,rey,wind250_1,lee250_1) 

ggplot() +
  stat_ma_line(df, mapping=aes(rey, lee250_1),  method="RMA",
               range.y = "relative", range.x = "relative", 
               linewidth = 1,fill = "yellow") +
  geom_point(df, mapping = aes(x = rey, lee250_1, colour=factor(ROS)), 
             size=3)+ 
  xlab("Re") + ylab((expression(paste(tau~"windward"))))+
  #scale_x_continuous(trans='log10') +
  #scale_y_continuous(trans='log10') +
  scale_color_manual(values = c("#0072B2", "#000000","#E7B800","#CC79A7")) + 
  labs(colour = "ROS (m/s)") +
  theme_bw()
#> Error: Discrete value supplied to continuous scale"

Peak fitting could be useful in some cases. New parameters n.min and n.max could be used to dynamically adjust the span, or the ignore_threshold.

This will require flipping of data but for bars also rotation of pairwise labels. angle aesthetic support in geom_label() and derived geoms is partly implemented in 'ggpp' 0.5.5 and will be implementrd in 'ggplot2' 3.5.5.

Dear Prof. Pedro, the example in the Readme is not working

I have the following message:

Warning messages:
1: In as.character.polynomial(polynom::as.polynomial(coefs), digits = coef.digits, :
NAs introduced by coercion

library(ggpmisc)

formula <- y ~ x + I(x^2)
ggplot(cars, aes(speed, dist)) +
  geom_point() +
  stat_fit_deviations(formula = formula, colour = "red") +
  stat_poly_line(formula = formula) +
  stat_poly_eq(aes(label =  paste(stat(eq.label), stat(adj.rr.label), sep = "*\", \"*")),
               formula = formula)

Thanks for the help

Add "staircase" pairwise contrasts.
Test with additional model fit functions.

stat_poly-eq fails with following message:
Warning messages:
1: In as.character.polynomial(polynom::as.polynomial(coefs), digits = coef.digits, :
NAs introduced by coercion
2: Computation failed in stat_poly_eq():
missing value where TRUE/FALSE needed
For testing purposes I used the following repex from Stack overflow https://stackoverflow.com/questions/7549694/add-regression-line-equation-and-r2-on-graph/35140066#35140066

library(ggplot2)
library(ggpmisc)
df <- data.frame(x = c(1:100))
df$y <- 2 + 3 * df$x + rnorm(100, sd = 40)
my.formula <- y ~ x
p <- ggplot(data = df, aes(x = x, y = y)) +

• geom_smooth(method = "lm", se=FALSE, color="black", formula = my.formula) +
• ggpmisc::stat_poly_eq(formula = my.formula,

•            aes(label = paste(..eq.label.., ..rr.label.., sep = "~~~")), 
  

•            parse = TRUE) +         
  

• geom_point()

p
Warning messages:
1: In as.character.polynomial(polynom::as.polynomial(coefs), digits = coef.digits, :
NAs introduced by coercion
2: Computation failed in stat_poly_eq():
missing value where TRUE/FALSE needed
No Text is output on graph

Using the example dataset in the help for stat_poly_eq...

set.seed(4321)
x <- 1:100
y <- (x + x^2 + x^3) + rnorm(length(x), mean = 0, sd = mean(x^3) / 4)
my.data <- data.frame(x = x, y = y,
                      group = c("A", "B"),
                      y2 = y * c(0.5,2),
                      w = sqrt(x))

# give a name to a formula
formula <- y ~ poly(x, 3, raw = TRUE)

I want to do something like the last example, in which the label is not the entire equation, but is constructed from the fitted parameters.

my.format <-
  "b[0]~`=`~%.3g*\", \"*b[1]~`=`~%.3g*\", \"*b[2]~`=`~%.3g*\", \"*b[3]~`=`~%.3g"
ggplot(my.data, aes(x, y)) +
  geom_point() +
  geom_smooth(method = "lm", formula = formula) +
  stat_poly_eq(formula = formula,
               output.type = "numeric",
               parse = TRUE,
               mapping =
                aes(label = sprintf(my.format,
                                    stat(coef.ls)[[1]][[1, "Estimate"]],
                                    stat(coef.ls)[[1]][[2, "Estimate"]],
                                    stat(coef.ls)[[1]][[3, "Estimate"]],
                                    stat(coef.ls)[[1]][[4, "Estimate"]])
                                    )
                   )

Except I want to use facets.

my.format <-
  "b[0]~`=`~%.3g*\", \"*b[1]~`=`~%.3g*\", \"*b[2]~`=`~%.3g*\", \"*b[3]~`=`~%.3g"
ggplot(my.data, aes(x, y)) +
  geom_point() +
  geom_smooth(method = "lm", formula = formula) +
  facet_wrap(~group) +
  stat_poly_eq(formula = formula,
               output.type = "numeric",
               parse = TRUE,
               mapping =
                aes(label = sprintf(my.format,
                                    stat(coef.ls)[[1]][[1, "Estimate"]],
                                    stat(coef.ls)[[1]][[2, "Estimate"]],
                                    stat(coef.ls)[[1]][[3, "Estimate"]],
                                    stat(coef.ls)[[1]][[4, "Estimate"]])
                                    )
                   )

The above does not work: the same coefficients are shown for both groups A and B.
However, when I try using geom="debug", the separate coefficients in coef.ls are returned correctly:

Input 'data' to 'draw_panel()':
                             Estimate   Std. Error    t value   Pr(>|t|)
(Intercept)             -2.219423e+04 3.293548e+04 -0.6738701 0.50376620
poly(x, 3, raw = TRUE)1  1.527045e+03 2.853578e+03  0.5351335 0.59513525
poly(x, 3, raw = TRUE)2 -1.478358e+01 6.631448e+01 -0.2229315 0.82457583
poly(x, 3, raw = TRUE)3  9.981851e-01 4.359348e-01  2.2897575 0.02667175

Input 'data' to 'draw_panel()':
                            Estimate   Std. Error    t value   Pr(>|t|)
(Intercept)             17289.934326 3.634737e+04  0.4756859 0.63654947
poly(x, 3, raw = TRUE)1   360.046135 3.055598e+03  0.1178316 0.90671408
poly(x, 3, raw = TRUE)2   -20.054699 6.924042e+01 -0.2896386 0.77339427
poly(x, 3, raw = TRUE)3     1.198387 4.464844e-01  2.6840515 0.01007669

Is this a bug, or am I doing something incorrectly?

According to my tests ggpmisc and gganimate are currently not compatible with each other, or please enlight me if I was wrong:

library(gganimate)
#> Loading required package: ggplot2
library(ggpmisc)
#> Loading required package: ggpp
#> 
#> Attaching package: 'ggpp'
#> The following object is masked from 'package:ggplot2':
#> 
#>     annotate

# Animation without stat_poly_eq()
ggplot(diamonds, aes(x = carat, y = price)) + geom_point() + geom_smooth(method = "lm") +
  transition_states(cut, transition_length = 1, state_length = 1) +
  enter_fade() + exit_shrink() +
  labs(title = "Cut = {closest_state}")
#> `geom_smooth()` using formula = 'y ~ x'

# Static plot with stat_poly_eq()
ggplot(diamonds, aes(x = carat, y = price)) + geom_point() + geom_smooth(method = "lm") +
    stat_poly_eq()
#> `geom_smooth()` using formula = 'y ~ x'

# Static plot with stat_poly_eq() and facets
ggplot(diamonds, aes(x = carat, y = price)) + geom_point() + geom_smooth(method = "lm") +
  stat_poly_eq() + facet_wrap(. ~ cut)
#> `geom_smooth()` using formula = 'y ~ x'

# Adding stat_poly_eq() to the animation causes the error
ggplot(diamonds, aes(x = carat, y = price)) + geom_point() + geom_smooth(method = "lm") +
  transition_states(cut, transition_length = 1, state_length = 1) +
  enter_fade() + exit_shrink() +
  stat_poly_eq() +
  labs(title = "Cut = {closest_state}")
#> `geom_smooth()` using formula = 'y ~ x'
#> Warning: Computation failed in `stat_poly_eq()`
#> Caused by error in `abs()`:
#> ! non-numeric argument to mathematical function
#> Error in `$<-.data.frame`(`*tmp*`, "group", value = ""): replacement has 1 row, data has 0

# Combining stat_poly_eq() with facets fails for each facet
ggplot(diamonds, aes(x = carat, y = price)) + geom_point() + geom_smooth(method = "lm") +
  transition_states(clarity, transition_length = 1, state_length = 1) +
  enter_fade() + exit_shrink() +
  stat_poly_eq() + facet_wrap(. ~ cut) +
  labs(title = "Clarity = {closest_state}")
#> `geom_smooth()` using formula = 'y ~ x'
#> Warning: Computation failed in `stat_poly_eq()`
#> Computation failed in `stat_poly_eq()`
#> Computation failed in `stat_poly_eq()`
#> Computation failed in `stat_poly_eq()`
#> Computation failed in `stat_poly_eq()`
#> Caused by error in `abs()`:
#> ! non-numeric argument to mathematical function
#> Error in `$<-.data.frame`(`*tmp*`, "group", value = ""): replacement has 1 row, data has 0

^{Created on 2023-06-18 with reprex v2.0.2}

It might also be the case that more efforts should be made by the author(s) of gganimte, but it would be nice if you could investigate a bit why they are not compatible

library(ggplot2)
library(ggpmisc)
library(tibble)

my_df <- data.frame(
  a = c(0, 1, 2, 7777),
  b = c('ggpmisc::geom_table cuts off text', 'one', 'two', 'seven thousand seven hundred and seventy seven')
)
my_tibble <- tibble(x = 0, y = 0, tbl = list(my_df))
t_geom_table <- ggplot(my_df) +
  geom_table(
    data    = my_tibble,
    mapping = aes(x = 0, y = 0, label = tbl),
    table.hjust    = 0,
    table.colnames = FALSE
  ) +
  theme_void()
t_geom_table

Here's an image of the output. The beginning of each line is also offset to the right, which may be related.

Using geom_debug() one sees that without groups or facets a data frame with two rows is returned. The example in the User Guide shows this. For some reason two versions of the same methods are run. The problem is not visible in the plot as two identical and overlapping tables are added to the plot with the top one hiding the lower one-

The Cran vignettes link has an error with the formatting of table 2, at least on chrome and edge browsers.
https://cran.r-project.org/web/packages/ggpmisc/vignettes/model-based-annotations.html

A question at Stackoverflow makes me think that the format string used for coeffient estimates could be a formal parameter of the stat_..._eq() functions.

Needs some tought as the formatting of the equation as a whole is far from trivial.

It needs to be explored if justification could be also adjusted for each text label dynamically in the position functions in the same way as the nudges are. This would be very useful when the length of the label text varies, as in this case constant nudge would result in varying distances.

The spacing of "text" output in stat_poly_eq() could be improved. Possibly, best is not to include the asterisk.

See question at Stackoverflow. The code below could work as a possible answer.

df <- data.frame(x = c(1:100))
df$y <- 20 + 30 * df$x + rnorm(100, sd = 80)

library(ggpmisc)
#> Loading required package: ggpp
#> Loading required package: ggplot2
#> 
#> Attaching package: 'ggpp'
#> The following object is masked from 'package:ggplot2':
#> 
#>     annotate

ggplot(df, aes(x, y)) + 
  geom_point() +
  stat_poly_line(method = "lm", se = FALSE) +
  stat_poly_eq(use_label(c("eq", "R2"), sep = ", "), output.type = "text") 

^{Created on 2022-09-06 with reprex v2.0.2}

'ggsignif' exports geom_signif() and stat_signif(), however pairwise comparisons are done individually, and p-values returned uncorrected. In addittion geom_signif() and stat_signif() are too much interdependent, making it impossible to use stat_signif() with other geometries. No labels to be parsed into expressions are returned and small p-values are not shown as < 0.001 or similar as they are in 'ggpmisc'.

After looking at 'ggsignif' code with the intention of submitting a pull request, it seems more reasonable to implement from scratch a statistic and a geometry with similar functionality in 'ggpmisc' than updating 'ggsignif'.

Getting this done soon and included in the 2nd edition of Learn R: As a Language would be ideal.

The geometry can be implemented first, and tested and used on its own, and later one or more statistics could be written making use of it: stat_fit_contrasts() and stat_multcomp().

The equivalents of use_label() for stat_fit_tb() to be used for selecting and renaming columns.
(Language translations of colum headers could be also implemented at a later time.)

Hi!

Thank you for the outstanding work on ggpmisc!

Over the past few days, I've been attempting, albeit unsuccessfully, to implement ggpmisc's stat_ma_eq and stat_ma_line within the upper plots of ggpairs. My objective is to display equations grouped by class in the upper plots with proper spacing, and scatterplots with fitted moving average (MA) lines (or any other type) at the bottom.

While I've managed to successfully integrate the lines, I've encountered challenges in creating a custom function for the upper plots. Specifically, I haven't found a method to ensure the proper display of equation overlap or the names of the corresponding taxa. Any guidance on resolving this issue would be greatly appreciated.

Perhaps a similar approach could be applied to other functions within ggpmisc.

Below is the code I've produced:

library(GGally)
library(ggplot2)
library(ggpmisc)


fun_name <– function(data, columns, grp_labels, color, ...) {
  log_data <- function(data) {
    data <- data %>% 
      mutate(across(where(is.numeric), ~ {log10(.)
      }))
    return(data)
  }
  
  upperfun <- function(data, mapping, grp.label){
    ggplot(data = data, mapping = mapping, grp.label=grp.label, color=color) +
      ggpmisc::stat_ma_eq(aes(label = after_stat(paste("bold(", {{grp.label}}, "*\":\")~~", 
                                                       eq.label, sep = ""))),
                          parse = TRUE) +
      geom_blank() +
      facet_wrap(~{{grp.label}}, ncol = 1, 
                 strip.position = "top")+ #I tried to pass them as facet to avoid overlap, but works poorly
      theme_void()+
      theme(strip.background = element_blank(),
            strip.text = element_blank(),
            panel.grid = element_blank())
  }
  
  lowerfun <- function(data, mapping){
    ggplot(data = data, mapping = mapping) +
      geom_point() +
      stat_ellipse(type = "norm") +
      ggpmisc::stat_ma_line(show.legend = T, se = FALSE, method = "lmodel2:MA") +
      theme(panel.grid = element_blank())
  }
  
  data <- log_data(data)
  
  data %>%
    ggpairs(columns = columns,
            ggplot2::aes(colour = {{color}}, alpha = 0.4),
            upper = list(continuous = wrap(upperfun, grp.label = grp_labels, ...)),
            lower = list(continuous = wrap(lowerfun, ...)),
            progress = FALSE) +
    theme_minimal(base_size = 9) +
    theme(panel.grid = element_blank())
}

Thank you

Best wishes
M.

The question at https://stackoverflow.com/questions/68998729/color-ggscatter-by-r-and-p-values is important. For stat_xxx_eq() this is possible for the equation, a similar approach applied to stat_xx_line() functions would allow coloring or even hiding fitted lines based on p.value of R^2. We should also make R^2adj available. This is fairly easy to implement...

Currently, the sep argument in use_label cannot take a \n for output.type=expression, since plotmath doesn't directly understand \n. However, there are ways to use, for example, atop to make multi-line expressions.

It would be great if a multi-line use_label functionality was added. One way this might work is by behind the scenes interpreting sep="\n" to trigger multiline atop functionality.

I just implemented function use_label() that uses paste() to assemble the combined label and supports a single value for sep. This is a simple user interface usable in many cases. A function fmt_label()could be also useful when users want to have arbitrary separators or additional text. This is to keep a note that if use_label() turns out to be limiting, a function using a template for the overall format would be a reasonable UI, although not necessarily very easy to implement, in particular handling of parsing errors.

Recent versions of 'ggplot2' have updated versions of statistics with formal parameter orientation. I have implemented this for the statistics in 'ggpmisc' except for these two. This would add consistency, and allow finding maxima and minima on the x values.

Please briefly describe your problem and what output you expect. If you have an ordinary question, please don't use this form, but instead ask on stackoverflow http://stackoverflow.com using tag [ggpmisc} and optionally [ggoplot2].

Please include a minimal reprex. The goal of a reprex is to make it as easy as possible for me to recreate your problem so that I can fix it. If you've never heard of a reprex before, start by reading https://www.tidyverse.org/help/#reprex, and follow the advice further down the page. Do NOT include session info unless it's explicitly asked for, or you've used reprex::reprex(..., si = TRUE) to hide it away.

Delete these instructions once you have read them.

With the ggpmisc 0.4.2 update, I'm now seeing this error pop up with existing code that worked with the previous release:

Error in after_stat(npcx) : could not find function "after_stat"

It appears that ggplot2::after_stat should be imported into the NAMESPACE.

Best,
Mike

These days I used your nice package, It looks brilliant. I got a minor question.

I want to save the values of peaks into a new file, but now i can only plot the values in the figure. Is there any way for me to do this?

Best,
Xu

When using the function symmetric_limits inside scale_y_continuous like this:

scale_y_continuous(limits = ggpmisc::symmetric_limits)

I get this result with the y-axis title label where it is not centered correctly. I have played around with hjust in theme for axis.title, but can't seem to get it to work. I know I am not providing reproducible code at the moment, and in my use case it is getting passed into the ggiraph package for an interactive plot. I'm not sure if this is fixable or not, or if you have any idea as to why this is happening. Thanks in advance for any help!

Running
ggpmisc 0.4.0
R version 4.0.3 (2020-10-10)
MacOS X 11.2.3 x86_64-apple-darwin17.0

library(tibble)
library(ggplot)
library(ggpmisc)

set.seed(0); tibble(a=1:100, b=runif(length(a), 0.7, 1.3) * (a*5)) %>% 
  ggplot(aes(a,b))+
  geom_point()+
  stat_poly_eq(aes(label=stat(eq.label)), formula=y~x, parse=T)


set.seed(0); tibble(a=1:100, b=runif(length(a), 0.7, 1.3) * (a*5)) %>% 
  ggplot(aes(a,b))+
  geom_point()+
  stat_poly_eq(aes(label=stat(eq.label)), formula=y~x+0, parse=T)

1. Running first chunk displays the equation y = -0.603 + 5.06 x
2. Running second chunk displays the equation y = 5.05 instead of y = 5.05 x

Add stat_sma_eq() and update stat_fit_deviations() and stat_fit_residuals().
Either package 'smatr' or 'lmodel2' could be used. Catch is that these packages only support very simple formulas, i.e., y ~ x which could surprise users is implemented in the same functions as 'lm', etc.

I suspect that 'smatr' would be the best option, but this needs to be further explored.

Support for ma and sma in stat_fit_deviations() would be particularly useful in teaching.

Would like to be able to show only the facets with high R squared. The code below does it but it would be easier if it could be done entirely within ggplot2, i.e. within the second pipeline. In particular we ran lm in the first pipeline and then again, implicitly, in the second pipeline. Also more important is having simpler code would be nice. A variation would be to show only the top k panels in R squared where k is specified.

library(broom)
library(dplyr)
library(ggplot2)
library(ggpmisc)

# find Trees that for which R squared >= 0.97.  Here all but tree 4.
Trees <- Orange %>%
  nest_by(Tree) %>%
  summarize(model = list(lm(age ~ circumference, data)), glance(model)) %>%
  filter(r.squared >= 0.97) %>%
  pull(Tree)

# plot
if (length(Trees)) {
  p <- Orange %>%
    filter(Tree %in% Trees) %>%
    ggplot(aes(circumference, age)) +
      geom_point() +
      stat_poly_eq() +
      geom_smooth(method = "lm", se = FALSE) +
      facet_wrap(~ Tree)
  plot(p)
 } 

I am trying to add a semi-transparent label to a plot using the geom_label_npc() function. Unfortunately, this does not seem to work, when I directly provide npcx, npcy and the label argument:

library(ggplot2); library(ggpmisc)

ggplot(data = mtcars) +
  geom_point(mapping = aes(wt, mpg)) +
  geom_label_npc(npcx=0.04, npcy=0.95, label="(a)", colour="white", alpha = 0.5) 

However, it seems to work, if I provide the same values inside a data.frame object:

df <- data.frame(
  x = 0.04,
  y = c(0.95),
  text = c("(a)")
)

ggplot(data = mtcars) +
  geom_point(mapping = aes(wt, mpg)) +
  geom_label_npc(data = df, aes(npcx = x, npcy = y, label = text), colour="white", alpha=0.5)

Is probably a small bug, that can easily be solved.

And another question, is there any way to use geom_label_npc() without the a border around the label or give the border a different colour to the text?

First of all, Thank you very much for your contribution. It looks brilliant.
I would like to ask you if there is any issue when installing your package on a Linux machine (Ubuntu 22.04) and R version 4.2.1.
I can not install the stable or unstable version. It says that splus2R library is not available and I cannot install it.
Could you please help with this?

ERROR: compilation failed for package ‘splus2R’

• removing ‘/home/guido/R/x86_64-pc-linux-gnu-library/4.2/splus2R’
  Warning in install.packages :
  installation of package ‘splus2R’ had non-zero exit status
  ERROR: dependency ‘splus2R’ is not available for package ‘ggpmisc’
• removing ‘/home/guido/R/x86_64-pc-linux-gnu-library/4.2/ggpmisc’
  Warning in install.packages :
  installation of package ‘ggpmisc’ had non-zero exit status

First of all, Thank you very much for your contribution. It looks brilliant.
I would like to ask you if there is any issue when installing your package in a mac machine.
I can not install the stable or unstable version.
Could you please help on this?

Error in loadNamespace(i, c(lib.loc, .libPaths()), versionCheck = vI[[i]]) :
namespace ‘quantreg’ 5.61 is being loaded, but >= 5.93 is required
Calls: ... withCallingHandlers -> loadNamespace -> namespaceImport -> loadNamespace
Execution halted
ERROR: lazy loading failed for package ‘ggpmisc’

Hi,
Thanks for your great tool, and I want to know how to extract the equations from stat_poly_eq, not just plot on the graph.

Best,
xiucz

In some cases, such as teaching it is useful to be able to swap x and y as response and explanatory variables. Support for this could be implemented in:
stat_poly_eq() and stat_quant_eq()
stat_fit_deviations() and stat_fit_residuals()
ggplot::stat_smooth() and ggplot::stat_quantiles()

see also #7

A statistic could be rather easily implemented by editing the code of stat_multcomp().

Need to create a reprex and investigate what is going on...

Looking at the documentation of stat_poly_eq and stat_quant_eq, do you think parse = [default] is important enough to appear in the arguments list, or is there a specific reason it doesn't appear there?

While I note it is discussed later, and the default is FALSE, that default is primarily for LaTeX users? Would a more common usage by others suggest a default of parse = TRUE might be preferred?

Some time ago a request/suggestion for supporting in stat_poly_eq() CIs for the fitted coefficients and other parameter estimates was made. More recently, a pull request at kassambara/ggpubr#418 and comments on it discussed CIs for correlation.

A CI for R2 could be useful in all the _eq statistics and could be obtained with package 'confintr'. For rank correlation CIs for tau and rho can be also computed using functions from 'confintr'. CIs for parameter estimates would be useful in stat_fit_tb() with summary tables. Could be implemented using confint() methods from 'stats', 'lme4' and 'MASS', or intervals() from 'nlme' .

This is a note to myself, for the future, only tentatively assigned to a milestone as most likely CIs will be implemented in stages at different milestones for different statistics. CIs for rho, tau and R^2 seem to be most useful and would help in achieving consistency in the available labels across similar statistics. Those for parameter estimates are in many circumstances more informative than t-values and p-values.

receiving error message with stat_poly_eq:

Warning message:
Computation failed in stat_poly_eq():
unable to find an inherited method for function ‘isa’ for signature ‘"formula"’

Would appreciate advice on this issue. Thank you.

Is it right that ggpmisc does not support lm models without interactions? Any easy way to handle it?

I suspect this is a problem in 'ggtext' rather than 'ggpmisc', but anyway spurious space is inserted where there is not any in eq.label.

Hi,

Thank you for a nice package!
Is there a way to determine the use of geom (either "text" or "label") based on whether the P value is significant or not?
I tried to do so but it didn't work.

x <- (1:100) / 10
y <- x + rnorm(length(x))
my.data <- data.frame(x = x,
                      y = y,
                      y.desc = - y,
                      group = c("A", "B"))


ggplot(my.data, aes(x, y)) +
  geom_point() +
  stat_correlation(geom = ifelse(after_stat(p.value) < 0.05, "text", "label"))

Thanks!

I'm running into an issue installing from source that is specific to v0.4.0. The previous v0.3.9 release installs correctly.

install.packages("https://cran.r-project.org/src/contrib/ggpmisc_0.4.0.tar.gz")

trying URL 'https://cran.r-project.org/src/contrib/ggpmisc_0.4.0.tar.gz'
Content type 'application/x-gzip' length 529683 bytes (517 KB)
==================================================
downloaded 517 KB

* installing *source* package ‘ggpmisc’ ...
** package ‘ggpmisc’ successfully unpacked and MD5 sums checked
** using staged installation
** R
** data
*** moving datasets to lazyload DB
** inst
** byte-compile and prepare package for lazy loading
Warning in .recacheSubclasses(def@className, def, env) :
  undefined subclass "numericVector" of class "Mnumeric"; definition not updated
Calls: <Anonymous> ... loadNamespace -> <Anonymous> -> .cacheClass -> .recacheSubclasses
** help
*** installing help indices
*** copying figures
** building package indices
** installing vignettes
** testing if installed package can be loaded from temporary location
Warning in .recacheSubclasses(def@className, def, env) :
  undefined subclass "numericVector" of class "Mnumeric"; definition not updated
Calls: <Anonymous> ... loadNamespace -> <Anonymous> -> .cacheClass -> .recacheSubclasses
** testing if installed package can be loaded from final location
Warning in .recacheSubclasses(def@className, def, env) :
  undefined subclass "numericVector" of class "Mnumeric"; definition not updated
Calls: <Anonymous> ... loadNamespace -> <Anonymous> -> .cacheClass -> .recacheSubclasses
** testing if installed package keeps a record of temporary installation path
* DONE (ggpmisc)

Here's my session info:

sessionInfo()

> sessionInfo()
R version 4.1.0 (2021-05-18)
Platform: x86_64-apple-darwin17.0 (64-bit)
Running under: macOS Big Sur 10.16

Matrix products: default
BLAS:   /Library/Frameworks/R.framework/Versions/4.1/Resources/lib/libRblas.dylib
LAPACK: /Library/Frameworks/R.framework/Versions/4.1/Resources/lib/libRlapack.dylib

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

other attached packages:
[1] AcidDevTools_0.3.14

loaded via a namespace (and not attached):
[1] compiler_4.1.0 tools_4.1.0

Hi, please, consider adding an instruction on how one can install ggpmisc package on an Ubuntu machine. I have faced the problem that clean Ubuntu is lacking from not straight-forward dependencies:

sudo apt update
sudo apt install gfortran liblapack-dev libopenblas-dev r-cran-sparsem

only after that we can run

install.packages("ggpmisc")

FYI, my sessionInfo():

R version 4.2.2 Patched (2022-11-10 r83330)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 22.04.2 LTS

Matrix products: default
BLAS:   /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3
LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/libopenblasp-r0.3.20.so

locale:
 [1] LC_CTYPE=C.UTF-8       LC_NUMERIC=C           LC_TIME=C.UTF-8
 [4] LC_COLLATE=C.UTF-8     LC_MONETARY=C.UTF-8    LC_MESSAGES=C.UTF-8
 [7] LC_PAPER=C.UTF-8       LC_NAME=C              LC_ADDRESS=C
[10] LC_TELEPHONE=C         LC_MEASUREMENT=C.UTF-8 LC_IDENTIFICATION=C

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

other attached packages:
[1] ggpmisc_0.5.2 ggpp_0.5.1    ggplot2_3.4.1

loaded via a namespace (and not attached):
 [1] magrittr_2.0.3     splines_4.2.2      MASS_7.3-58.2      tidyselect_1.2.0
 [5] munsell_0.5.0      colorspace_2.1-0   lattice_0.20-45    R6_2.5.1
 [9] rlang_1.0.6        fansi_1.0.4        dplyr_1.1.0        tools_4.2.2
[13] grid_4.2.2         gtable_0.3.1       quantreg_5.94      utf8_1.2.3
[17] cli_3.6.0          withr_2.5.0        MatrixModels_0.5-1 survival_3.5-3
[21] tibble_3.1.8       lifecycle_1.0.3    Matrix_1.5-1       vctrs_0.5.2
[25] glue_1.6.2         compiler_4.2.2     pillar_1.8.1       generics_0.1.3
[29] scales_1.2.1       SparseM_1.81       pkgconfig_2.0.3

Hi @aphalo,

many thanks for developing this very helpful package!

I just wanted to make you aware of a very small typo (duplicated punctuation) in the ggpmisc manual that I just stumbled across when trying to add p.values using stat_poly_eq with aes(label = paste(..p.value..label..).

As one would guess, the computed variable name is p.value.label not p.value..label.

Have a nice day!

Make stat_poly_eq(), stat_quant_eq() and stat_ma_eq() fully respect R option options(OutDec = ","). It seems the model equation respects it but not R2 and possibly other labels. See https://stackoverflow.com/questions/74813032/modifying-the-decimal-mark-of-stat-poly-eq

I based stat_quant_line() partly on ggplot2::stat_quantile() while I wrote stat_quant_eq() almost from scratch. The current code for stat_quant_line() makes it impossible to retrieve the formula and class from the fitted model. As I have made it possible to use methods that modify the formula or call user defined functiions that in turn call a model fit function returning an acceptable fited model object, this represents a problematic inconsistency.

A rewrite based in stat_quant_eq() should not be too difficult.