schmidtpaul/biomathr

Utility functions used at BioMath

BioMathR

The R package {BioMathR} provides a collection of utility and wrapper functions, {ggplot2} themes and color palettes, all of which are often used at BioMath.

Installation

You can install BioMathR quickly like so:

remotes::install_github('SchmidtPaul/BioMathR')

Or you can install it with all its suggested packages (i.e. extra packages that may only be necessary for one of the functions in BioMathR) like so:

remotes::install_github("SchmidtPaul/BioMathR", dependencies = TRUE, suggest = TRUE)

Find the documentation of all functions by clicking on the "Reference" tab at the top.

biomathr's Issues

Optionally use `fwrite()` instead of `write.table()` for `write_ascii()`

To improve speed. May work like so:

data.table::fwrite(
  x = data,
  file = path,
  sep = "\t",
  dec = dec,
  col.names = TRUE,
  row.names = FALSE,
  quote = FALSE,
  fileEncoding = "UTF-8"
)

improve `describe()`

• It cannot handle group_by(name), because it is also creating a name column

Add `get_desc()`

Base on dlookr::describe(weight)

• make sure not to load the package, as it loads {showtext} which leads to rendering issues with {ggplot2}
• optional German
• optional print-ready table

Add `get_residual_plots()`

Base on this:

residuencheck <-
  function(model,
           col_dots = BMcols["grey"],
           col_line = BMcols["green"]) {
    
    # Varianzhomogenität
    rp1 <- plot(check_heteroscedasticity(model)) + theme_BioMath()
    rp1$layers[[1]]$aes_params$colour <- col_dots
    rp1$layers[[2]]$aes_params$colour <- col_line
    rp1$layers[[2]]$aes_params$fill   <- col_line
    
    # Normalverteilung
    rp2 <- plot(check_normality(model), type = "qq") + theme_BioMath()
    rp2$layers[[1]]$aes_params$fill   <- col_line
    rp2$layers[[2]]$aes_params$colour <- col_dots
    rp2$layers[[3]]$aes_params$colour <- col_line
    
    plot_grid(rp1, rp2, nrow = 1)
  }

• replace {cowplot} with {patchwork}
• optional test results?
• optional German?

Handle multiple yvars differently in `desc_tabs()`

They should not be coloured with the same conditional formatting, since their scale may differ a lot.

Suppress warning in `desplot_across()`

Warning shows up as:

Warnmeldung:
In abbreviate(lt.text, 2, method = "both") :
  abbreviate mit Nicht-ASCII-Zeichen genutzt

add `gg_export(fast = TRUE)`

Via suggesting {ragg} as device

https://statistik-dresden.de/archives/17146?utm_source=ReviveOldPost&utm_medium=social&utm_campaign=ReviveOldPost&amp

Add `get_varcomp()`

Base on this:

get_vc <- function(model) {
  model %>%
    VarCorr() %>%
    as_tibble() %>%
    setNames(c('group', 'effect', 'eff_2', 'var', 'sd')) %>%
    mutate(
      var_prop = var / sum(var),
      var_p = case_when(
        var_prop < 0.000001 ~ "   0%",
        TRUE ~ percent(var_prop, accuracy = 0.1, trim = FALSE)
      ),
      effect   = gsub(effect, pattern = '[\\(,\\)]', replacement = ''),
      effect   = ifelse(is.na(effect), '', effect)
    ) %>%
    select(group, var, var_p, sd, effect, everything())
}

• which model classes are supported?
• optional German?
• optional print-ready version?

Add `desplot_BM()`

Base on this:

fargs <- formals(desplot)
fargs$flip <- TRUE
fargs$cex <- 1
fargs$out1.gpar <- list(col = "black", lwd = 1, lty = "dashed")
fargs$out2.gpar <- list(col = "black", lwd = 1, lty = "dotted")
formals(desplot) <- fargs

Add `project_setup()`

• create folder structure
• 00 setup.R
  • create
• 01 import.R
  • create
  • source(here::here("10 RCode" , "code", "00 setup.R"), encoding = "UTF-8")