thomas-neitmann/ggcharts

Get You to Your Desired Plot Faster

Hi there 👋

Some facts about myself:

me <- list(
  name = "Neitmann",
  first_name = "Thomas",
  occupation = "Associate Director, Data Science",
  employer = "Denali Therapeutics",
  favorite_programming_language = "R",
  my_r_packages = c("admiral", "ggcharts", "mdthemes"),
  blog = "https://thomasadventure.blog/"
)

Here are my top 5 most popular R blog posts:

• The Comprehensive Guide to Installing R Packages from CRAN, Bioconductor, GitHub and Co.
• Understanding Non-Standard Evaluation. Part 1: The Basics
• All You Need To Know About Merging (Joining) Datasets in R
• Transform a {ggplot2} Axis to a Percentage Scale
• Adding Labels to a {ggplot2} Bar Chart

ggcharts's Issues

Error in (function (el, elname) :

When I try to use ggcharts, encounter Error

here is the code and error message：
base_plot <- ggplot(mtcars, aes(hp, mpg)) + geom_point(color = "gray")
p1 <- base_plot + theme_ggcharts()

Error in (function (el, elname) :
"plot.title.position" is not a valid theme element name.

Issue with faceted `bar_chart()` when sort = FALSE

library(ggcharts)
bar_chart(biomedicalrevenue, year, revenue, facet = company, sort = FALSE, horizontal = FALSE)

bar_chart() sorts when sort = FALSE

library(ggcharts)
data(biomedicalrevenue)
revenue_2018 <- biomedicalrevenue[biomedicalrevenue$year == 2018, ]
bar_chart(revenue_2018[c(3, 9, 2, 20, 13), ], company, revenue, sort = FALSE)

Create `line_chart()` function

Create hex sticker for {ggcharts}

Make `dumbbell_chart()` work with data in long format

Currently, dumbbell_chart() only works with data in wide format. Add code to handle data in long format.

Create `likert_chart()` function

The function should be able to create plots like this one.

Source: http://rnotr.com/likert/ggplot/barometer/likert-plots/

Document `revenue_wide` dataset

Create `regression_chart()` function

Create a function to easily create plots of linear regression models.

Add `highlight` feature to `diverging_*_chart()` functions

Unlike the highlight feature in bar_chart and lollipop_chart, the non-highlighted bars/lollipops should not be lightgray but a tint of the highlight color.

Don't display legend title in `line_chart()`

Don't display axis label for categorical axes

This affects bar_chart(), lollipop_chart() and dumbbell_chart().

Create `scatter_chart()`

Add palettes for themes

Use Google fonts in {ggcharts} themes

Warning: `expand_scale()` is deprecated; use `expansion()` instead.

Update code to avoid this error getting printed with the latest version of {ggplot2}.

Modify bar_chart {ggcharts} Example

Modify/Change bar_chartexample from:
biomedicalrevenue %>%
filter(year == 2018) %>%
bar_chart(company, year, limit = 10)
to a potentially more pleasing:
...
bar_chart(company, revenue, limit = 10)

Legend pops up when changing theme

Currently, the highlight feature is implement via scale_fill_manual(). This has the consequence that when changing the theme a legend pops up. Implementing the highlight feature via scall_fill_identity() should fix this.

plot <- biomedicalrevenue %>%
  filter(year == 2014) %>%
  bar_chart(company, revenue, limit = 10, highlight = "Sanofi") %>%
  print()

plot + theme_classic()

Display "other" in chart

When using the threshold argument of bar_chart() and lollipop_chart() it would be nice to have an option to display a bar/lollipop at the bottom that sums up the values of all categories falling below the threshold.

data <- data.frame(
  cat = letters[1:10],
  val = c(4, 5, 2, 1, 22, 12, 32, 11, 5, 1),
  stringsAsFactors = FALSE
)

reorder_other <- function(x, by) {
  x2 <- x[x != "other"]
  by2 <- by[x != "other"]
  levels <- c("other", x2[order(by2)]) %>% print()
  ordered(x, levels = levels)
}

p <- data %>%
  mutate(cat2 = ifelse(val > 4, cat, "other")) %>%
  group_by(cat2) %>%
  summarise(val = sum(val)) %>%
  ungroup() %>%
  mutate(cat2 = reorder_other(cat2, val)) %>%
  bar_chart(cat2, val, sort = FALSE) %>%
  print()

In addition an automatic footnote should be generated.

p + labs(caption = "'Other' contains all categories with a value less than or equal to 4.")

Replace `stop()` with `rlang::abort()`

All errors should be thrown with rlang::abort() instead of stop().

Add `label` parameter to `bar_chart()`

Add `stacked_bar_chart` function

Dynamic color for charts using fill

Do you have examples with the usage for fill?

I have the following code:

items %>%
  dplyr::group_by(state) %>%
  dplyr::summarise(n = sum(n)) %>%
  dplyr::mutate(col = ifelse(n > 50, "#1F77B4", "#2CA02C")) %>%
  ggcharts::bar_chart(
    state, 
    n, 
    limit = 15,
    fill = col
  )

That produces the following plot:

If I add + scale_fill_identity() it then produces:

So I wonder what's the correct usage when specifying the fill attribute for ggplot.

Refactor `diverging_*_chart()` functions

Although diverging_bar_chart() and diverging_lollipop_chart() have almost the same code, they currently live separated in two files. That way whenever something needs to be updated, it has to been done in both files which is cumbersome and error prone.

To mitigate this, create a diverging_chart() function that is called internally by diverging_bar_chart() and diverging_lollipop_chart(). This function should do all the work.

Add `pyramid_chart()` function

A simple sequence of the data frame should be sufficient to achieve the order sequence rather then pulling out the variable and sequencing on it.

I noticed that in the pyramid_chart function if there is no sorting order then you create a sequence vector and assign it to order.
order <- data %>% dplyr::pull(!!x) %>% seq_along()
It would be a little efficient to just to use order <- 1:nrow(data).
Is there a specific reason why you do this by pulling the variable out?
I get that for creating the limit you have to pull the y column out but I don't think you'd need to do that for order.

pyramid_chart lost half tag

Hi guys,
I really appreciate your ggcharts package.
But a bug occur to me.
When I try pyramid_chart with geom_text(), it lost the left half tagg.

Here is the code：
data(popch) options(scipen = 999) pyramid_chart(popch, age, pop, sex, xlab = "Population", title = "Switzerland 2020") + geom_text(aes(label=round(pop/10000,2)))

Add `column_chart()` function

The column_chart() function should be a wrapper of bar_chart() with horizontal = FALSE and sort = !is.numeric(x) as default.

Issue with faceted `bar_chart()` when horizontal = FALSE

library(ggcharts)
data(biomedicalrevenue)
bar_chart(biomedicalrevenue, company, revenue, facet = year, limit = 3, horizontal = FALSE)

Setup `lintr`

Automatically attach ggplot2 when calling library(ggcharts)

Add ggplot2 to the Depends field in DESCRIPTION to achieve this.

Create shiny gadgets to make creating charts even easier

Add population dataset for `pyramid_chart()` examples

Rename `limit` parameter to `top_n`

Update documentation to include more examples

Currently, the examples only show the very basic usage of the *_chart() functions. The following features deserve their own examples:

• facet
• highlight
• sort
• *_color

Heat Map, Histogram and Box Plot

Kindly add Heat Map, Histogram, and Box-Plot. It will be very helpfull. This is a wonderful package.

Add `dumbbell_chart()` function

Adjust default theme

Adjust the default theme to include a proper x axis line with ticks, greater font size and possibly a different default font. Think about removing all grid lines.

Setup continuous deployment for pkgdown website

Remove the tidytext dependency

Currently, {ggcharts} depends on {tidytext} because it uses reorder_within() and scale_x_reordered(). Both functions are super short and require only the {base} package and {ggplot2}, whereas the rest of {tidytext} has lots of dependencies. Integrate the two functions into {ggcharts} source code to avoid having to install all these dependencies.

Create vignette on how to customize `ggcharts` with `ggplot2` functions

Add `histogram()` function

`bar_chart()` and `lollipop_chart()` should handle missing `y` argument

Similar to geom_bar(), in case y is missing the count of each unique x value should be displayed.

Highlight Vignette

Create a vignette showcasing the highlight feature of bar_chart() mad lollipop_chart() .

Standard evaluation functions

All functions currently use non-standard evaluation which is great for interactive use but IMO clumsy to program with. Having functions that accepts strings as arguments could come in very handy in {shiny} apps.

Double x axis labels when sort = FALSE and horizontal = FALSE

When calling bar_chart() with sort = FALSE and horizontal = FALSE each x axis label appears twice in the plot:

library(dplyr)
library(ggcharts)
data(biomedicalrevenue)

biomedicalrevenue %>%
  filter(company %in% c("Roche", "Novartis")) %>%
  bar_chart(year, revenue, facet = company, sort = FALSE, horizontal = FALSE)

When horizontal = TRUE this does not occur:

biomedicalrevenue %>%
  filter(company %in% c("Roche", "Novartis")) %>%
  bar_chart(year, revenue, facet = company, sort = FALSE, horizontal = FALSE)

Add gallery to README

Add `devtools::spell_check()` to Travis CI pipeline

Add threshold argument to bar_chart() and lollipop_chart()

In addition to the limit argument, which can be used to display the top n categories, add a threshold argument. Only categories with values above the threshold should be displayed in the chart.

Switch CI to GitHub actions