In Tony-Chen-Melbourne/bcggtheme: Create ggplot charts in the style of BCG

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.height = 4,
  fig.width = 6,
  dev = "cairo_pdf"
)

library(tidyverse)
library(bcggtheme)
library(unikn)
library(scales)

Using the package

This vignette explains how to use bcggtheme to apply BCG-style chart formatting to charts made in R using ggplot.

Download bcggtheme using

devtools::install_github("Tony-Chen-Melbourne/bcggtheme")

After downloading the package, there is one more step required before it will work. The base font in the package is MS Trebuchet, which needs to be loaded into R. This can be done using the extrafont package.

install.packages(extrafont) extrafont::font_import()

This will take a few minutes to do its thing, after which you're good to go. The import only needs to be done once, after which bcggtheme takes care of the rest.

bcggtheme allows for two different chart styles: classic and modern.

BCG Classic

For example, using the in-built iris dataset:

plot <- ggplot(iris,
               aes(x = Sepal.Length,
                   y = Sepal.Width,
                   colour = Species)) +
        geom_point(size = 4) +
        labs(x = "Species",
             y = "",
             colour = "Species")

plot

Then we can add the overlaying theme, using bcg_theme_classic. This changes the font, font size, axes, and labels.

plot +
  bcg_theme_classic()

Then change the y scale with bcg_scale_y_continuous(). This gets rid of the gap where the y axis joins the x axis:

plot +
  bcg_theme_classic() +
  bcg_scale_y_continuous()

Finally, adjust the colours using bcg_colour_manual:

plot +
  bcg_theme_classic() +
  bcg_scale_y_continuous() +
  bcg_colour_manual()

The standard background colour for bcggtheme is white. In order to get the background the same grey colour as a BCG key message, use bcg_theme_classic(background = "grey)

plot +
  bcg_theme_classic(background = "grey") +
  bcg_scale_y_continuous() +
  bcg_colour_manual()

BCG Modern

This works best for bar charts, as will be explained later. Starting with a made-up data set:

data <- tribble(~person, ~citations, ~school,
                "Summers", 156191, "Harvard",
                "Duflo", 65808, "MIT",
                "Pakes", 39414, "Harvard",) %>% 
  mutate(person = fct_reorder(person, citations))

plot <- ggplot(data,
               aes(x = person,
                   y = citations,
                   fill = school)) +
        geom_col() +
        labs(x = "Economist",
             y = "",
             subtitle = "Citations",
             colour = "School")

plot

Then we can add the overlaying theme, using bcg_theme_modern. The default option removes the y axis, and removes all axis tick marks

plot +
  bcg_theme_modern(legend = "right")

Then we can add the other bells-and-whistles as before, noting that the relevant colour function for is bcg_fill_manual.

plot +
  bcg_theme_modern(legend = "right") +
  bcg_scale_y_continuous() +
  bcg_fill_manual()

Clearly this chart isn't very useful, because we don't know anything about the actual magnitude of citations for each economist. The way around this is to add in labels. We do this using bcg_geom_label, which adds standard grey box labels at the top of each bar. You will then have to use nudge_y to move the box up a little bit.

plot +
  bcg_theme_modern(legend = "right") +
  bcg_scale_y_continuous(limits = c(0,200000)) +
  bcg_fill_manual() +
  bcg_geom_label(aes(label = scales::comma(citations)), nudge_y = 20000)

With the modern version we can add the y axis line back in, using bcg_theme_modern(y_axis = TRUE).

plot +
  bcg_theme_modern(legend = "right", y_axis = TRUE) +
  bcg_scale_y_continuous(limits = c(0,200000)) +
  bcg_fill_manual() +
  bcg_geom_label(aes(label = scales::comma(citations)), nudge_y = 20000)

Flipped charts

bcggtheme has the function to remove the x, so that charts make sense when used in conjunction with coord_flip().

plot +
  bcg_theme_modern(legend = "right", y_axis = TRUE, background = "grey") +
  bcg_scale_y_continuous(limits = c(0,200000)) +
  bcg_fill_manual() +
  bcg_geom_label(aes(label = scales::comma(citations)), nudge_y = 20000) +
  coord_flip()

Then setting bcg_theme_modern(flipped = TRUE)

plot +
  bcg_theme_modern(legend = "right",
                   y_axis = TRUE,
                   background = "grey",
                   flipped = TRUE) +
  bcg_scale_y_continuous(limits = c(0,200000)) +
  bcg_fill_manual() +
  bcg_geom_label(aes(label = scales::comma(citations)), nudge_y = 20000) +
  coord_flip()

Note that the same functionality applies to bcg_theme_classic().

Colours

The package is loaded with most standard BCG colours. Colours and palettes are tedious to add, so it's a work in progress and some may be missing. The colours are divided across three palettes: bright, dark, ee, multi, traffic (picking colours from left to right).

bcg_colour_manual and bcg_fill_manual can any choice by specifying pal = "X". The default option is pal = bright. The colouring order can also be reversed using the reverse = TRUE.

### Outputting charts from R There are two functions to output charts: `bcg_save` and `bcg_save_pptx`. They automatically save the last object created through ggplot. Both output charts in up to 5 formats: - "third": 11 cm width $\times$ 14cm height - "half": 14 cm width $\times$ 14cm height - "two third": 18 cm width $\times$ 14cm height - "large": 24 cm width $\times$ 14cm height - "full": 30 cm width $\times$ 14cm height These sizes are designed to replicate the options available in ThinkCell. `bcg_save` and `bcg_save_pptx` can save any one of these formats through `type = "chosen format"`. Alternatively, all five formats can be saved by setting `type = "all"`. For example, the following code would create a folder called `test` within the working directory, with the relevant png charts contained within wzxhzdk:15 The following code would create a folder called `test` within the working directory, with the relevant pptx charts contained within. The powerpoint slides contain editable graphics, which can be easily copied across to your desired presentation. wzxhzdk:16 `bcg_save` works best with `.png` files, currently it is unable to output `.pdf` files. (This shouldn't be a problem. In this line of work, I cannot imagine why you would want to directly output to pdf). I recommend that `bcg_save_pptx()` shouldn't be used with map outputs, because the editable powerpoint graphic can easily blow up in size (a simple map of Australia can easily go to 1mb or more) and become unwieldy.

Tony-Chen-Melbourne/bcggtheme documentation built on Oct. 23, 2020, 4:57 p.m.