Introduction to csstyle

library(csstyle)
#> csstyle 2026.8.21
#> https://niphr.github.io/csstyle/
library(ggplot2)

Overview

csstyle standardizes graphs, tables, and reports to follow the Core Surveillance visual guidelines. The design is deliberately narrow. The package does not expose every ggplot2 option. It offers a small set of outputs that look the same, whoever produces them and whenever.

The package covers four areas.

ggplot2 themes and color scales
Color palettes defined centrally and applied consistently
Number and date formatting in both Norwegian and international journal styles
Small utility functions for common axis tasks

Themes

theme_cs() is the main theme function. It applies the Core Surveillance appearance to any ggplot2 plot.

# Basic scatter plot with Core Surveillance theme
ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point() +
  theme_cs() +
  labs(
    title = "Engine Power vs Fuel Efficiency",
    x = "Miles per Gallon",
    y = "Horsepower"
  )

A few arguments adjust the layout without breaking the overall style.

# Theme with bottom legend and vertical x-axis labels
ggplot(mtcars, aes(x = factor(cyl), y = mpg, fill = factor(cyl))) +
  geom_boxplot() +
  theme_cs(legend_position = "bottom", x_axis_vertical = TRUE) +
  labs(title = "MPG by Number of Cylinders", fill = "Cylinders")

Color palettes

The package ships several predefined palettes. Inspect them directly.

# View available colors
head(colors$named_colors)
#>        H1        H2        H3        H4        H5        H6 
#> "#393C61" "#0975B5" "#2EA1C0" "#709900" "#B11643" "#FC5F56"

# Display all palettes
display_all_palettes()

Apply a palette to a plot with scale_color_cs() or scale_fill_cs().

# Using the primary color palette
ggplot(mtcars, aes(x = mpg, y = hp, color = factor(cyl))) +
  geom_point(size = 3) +
  scale_color_cs(palette = "primary") +
  theme_cs() +
  labs(color = "Cylinders")

# Using the warning palette for fills
ggplot(mtcars, aes(x = factor(cyl), fill = factor(cyl))) +
  geom_bar() +
  scale_fill_cs(palette = "warning") +
  theme_cs() +
  labs(title = "Car Count by Cylinders", fill = "Cylinders")

Number formatting

Every number formatting function belongs to one of two parallel families. The two families reflect the intended audience. The Norwegian family follows domestic conventions. The journal family follows international conventions for academic publications. The “Journal formatting” section below compares the two side by side.

Norwegian number format

The format_num_as_nor_* functions use a comma for the decimal separator, a space for thousands, and "IK" (ikke kjent) for missing values.

# Format numbers with Norwegian conventions
numbers <- c(1234.56, 9876.54, 123.45, NA)

# Basic number formatting (0, 1, 2 decimal places)
format_num_as_nor_num_0(numbers)
#> [1] "1235" "9877" "123"  "IK"
format_num_as_nor_num_1(numbers)
#> [1] "1234,6" "9876,5" "123,5"  "IK"
format_num_as_nor_num_2(numbers)
#> [1] "1234,56" "9876,54" "123,45"  "IK"

# Percentage formatting
percentages <- c(12.34, 56.78, 90.12)
format_num_as_nor_perc_1(percentages)
#> [1] "12,3 %" "56,8 %" "90,1 %"

# Per 100k population rates
rates <- c(123.45, 678.90)
format_num_as_nor_per100k_1(rates)
#> [1] "123,5 /100k" "678,9 /100k"

Date formatting

Norwegian date format

The format_date_as_nor() and format_datetime_as_nor() functions produce dates in the Norwegian dd.mm.yyyy style.

# Current date
format_date_as_nor()
#> [1] "26.09.2026"

# Specific dates
test_date <- as.Date("2023-12-25")
format_date_as_nor(test_date)
#> [1] "25.12.2023"

# Datetime formatting
test_datetime <- as.POSIXct("2023-12-25 14:30:00")
format_datetime_as_nor(test_datetime)
#> [1] "25.12.2023 kl. 14:00"

# Filename-safe datetime
format_datetime_as_file(test_datetime)
#> [1] "2023_12_25_143000"

Journal formatting

For academic publications the package provides a parallel set of functions. They use a decimal point, a comma thousands separator, "NA" for missing values, and ISO 8601 dates.

Number format comparison

# Compare Norwegian vs Journal formatting
numbers <- c(1234.56, 9876.54, NA)

# Norwegian format (space thousands, comma decimal, "IK" for NA)
format_num_as_nor_num_1(numbers)
#> [1] "1234,6" "9876,5" "IK"

# Journal format (comma thousands, decimal point, "NA" for NA)
format_num_as_journal_num_1(numbers)
#> [1] "1,234.6" "9,876.5" "NA"

# Percentage comparison
percentages <- c(12.34, 56.78)

# Norwegian: "12,3 %" vs Journal: "12.3%"
format_num_as_nor_perc_1(percentages)
#> [1] "12,3 %" "56,8 %"
format_num_as_journal_perc_1(percentages)
#> [1] "12.3%" "56.8%"

# Per 100k comparison  
rates <- c(123.45, 678.90)

# Norwegian: "123,5 /100k" vs Journal: "123.5/100k"
format_num_as_nor_per100k_1(rates)
#> [1] "123,5 /100k" "678,9 /100k"
format_num_as_journal_per100k_1(rates)
#> [1] "123.5/100k" "678.9/100k"

Date format comparison

test_date <- as.Date("2023-12-25")
test_datetime <- as.POSIXct("2023-12-25 14:30:00")

# Norwegian format: "25.12.2023"
format_date_as_nor(test_date)
#> [1] "25.12.2023"

# Journal format (ISO 8601): "2023-12-25"
format_date_as_journal(test_date)
#> [1] "2023-12-25"

# Datetime comparison
format_datetime_as_nor(test_datetime)     # "25.12.2023 kl. 14:00"
#> [1] "25.12.2023 kl. 14:00"
format_datetime_as_journal(test_datetime)  # "2023-12-25 14:30:00"
#> [1] "2023-12-25 14:30:00"

Log scale transformations

Both families include inverse log transformation variants. Use them when the axis values are on a log scale and you want the labels to show the original units.

log_values <- c(1, 2, 3)

# Log2 transformations (2^1, 2^2, 2^3 = 2, 4, 8)
format_num_as_nor_invlog2_1(log_values)      # "2,0", "4,0", "8,0"
#> [1] "2,0" "4,0" "8,0"
format_num_as_journal_invlog2_1(log_values)  # "2.0", "4.0", "8.0"
#> [1] "2.0" "4.0" "8.0"

# Log10 transformations (10^1, 10^2, 10^3 = 10, 100, 1000)  
format_num_as_journal_invlog10_1(log_values) # "10.0", "100.0", "1,000.0"
#> [1] "10.0"    "100.0"   "1,000.0"

Utility functions

Pretty breaks

pretty_breaks() generates axis break positions that round to human-friendly values.

ggplot(mtcars, aes(x = mpg, y = hp)) +
  geom_point() +
  scale_y_continuous(breaks = pretty_breaks(n = 4)) +
  theme_cs()

Every nth label

On crowded discrete axes, every_nth() keeps only every nth tick label and drops the rest.

ggplot(mtcars, aes(x = rownames(mtcars), y = mpg)) +
  geom_col() +
  scale_x_discrete(breaks = every_nth(n = 4)) +
  theme_cs() +
  set_x_axis_vertical()