Technology Tales

How to centre titles, remove gridlines and write reusable functions in {ggplot2}

Published on 20^th March 2026 Estimated Reading Time: 9 minutes

{ggplot2} is widely used for data visualisation in R because it offers a flexible, layered grammar for constructing charts. A plot can begin with a straightforward mapping of data to axes and then be refined with titles, themes and annotations until it better serves the message being communicated. That flexibility is one of the greatest strengths of {ggplot2}, though it also means that many useful adjustments are small, specific techniques that are easy to overlook when first learning the package.

Three of those techniques fit together particularly well. The first is centring a plot title, a common formatting need because {ggplot2} titles are left-aligned by default. The second is removing grid lines and background elements to produce a cleaner, less cluttered appearance. The third is wrapping familiar {ggplot2} code into a reusable function so that the same visual style can be applied across different datasets without rewriting everything each time. Together, these approaches show how a basic plot can move from a default graphic to something more polished and more efficient to reproduce.

Centring the Plot Title

A clear starting point comes from a short tutorial by Luis Serra at Ubiqum Code Academy, published on RPubs, which focuses on one specific goal: centring the title of a {ggplot2} output. The example uses the well-known Iris dataset, which is included with R and contains 150 observations across five variables. Those variables are Sepal.Length, Sepal.Width, Petal.Length, Petal.Width and Species, with Species stored as a factor containing three levels (setosa, versicolor and virginica), each represented by 50 samples.

The first step is to load {ggplot2} and inspect the structure of the data using library(ggplot2), followed by data("iris") and str(iris). The structure output confirms that the first four columns are numeric, and the fifth is categorical. That distinction matters because it makes the dataset well suited to a scatter plot with a colour grouping, allowing two continuous variables to be compared while species differences are shown visually.

The initial chart plots petal length against petal width, with points coloured by species:

ggplot() + geom_point(data = iris, aes(x = Petal.Width, y = Petal.Length, color = Species))

This produces a simple scatter plot and serves as the base for later refinements. Even in this minimal form, the grammar is clear: the data are supplied to geom_point(), the x and y aesthetics are mapped to Petal.Width and Petal.Length, and colour is mapped to Species.

Once the scatter plot is in place, a title is added using ggtitle("My dope plot"), appended to the existing plotting code. This creates a title above the graphic, but it remains left-justified by default. That alignment is not necessarily wrong, as left-aligned titles work well in many visual contexts, yet there are situations where a centred title gives a more balanced appearance, particularly for standalone blog images, presentation slides or teaching examples.

The adjustment required is small and direct. {ggplot2} allows title styling through its theme system, and horizontal justification for the title is controlled through plot.title = element_text(hjust = 0.5). Setting hjust to 0.5 centres the title within the plot area, whilst 0 aligns it to the left and 1 to the right. The revised code becomes:

ggplot() +
  geom_point(data = iris, aes(x = Petal.Width, y = Petal.Length, color = Species)) +
  ggtitle("My dope plot") +
  theme(plot.title = element_text(hjust = 0.5))

That small example also opens the door to a broader understanding of {ggplot2} themes. Titles, text size, panel borders, grid lines and background fills are all managed through the same theming system, which means that once one element is adjusted, others can be modified in a similar way.

Removing Grids and Background Elements

A second set of techniques, demonstrated by Felix Fan in a concise tutorial on his personal site, begins by generating simple data rather than using a built-in dataset. The code creates a sequence from 1 to 20 with a <- seq(1, 20), calculates the fourth root with b <- a^0.25 and combines both into a data frame using df <- as.data.frame(cbind(a, b)). The plot is then created as a reusable object:

myplot = ggplot(df, aes(x = a, y = b)) + geom_point()

From there, several styling approaches become available. One of the quickest is theme_bw(), which removes the default grey background and replaces it with a cleaner black-and-white theme. This does not strip the graphic down completely, but it does provide a more neutral base and is often a practical shortcut when the standard {ggplot2} appearance feels too heavy.

More selective adjustments can also be made independently. Grid lines can be removed with the following:

theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank())

This suppresses both major and minor grid lines, whilst leaving other parts of the panel unchanged. Borderlines can be removed separately with theme(panel.border = element_blank()), though that does not affect the background colour or the grid. Likewise, the panel background can be cleared with theme(panel.background = element_blank()), which removes the panel fill and borderlines but leaves grid lines in place. Each of these commands targets a different component, so they can be combined depending on the desired result.

If the background and border are removed, axis lines can be added back for clarity using theme(axis.line = element_line(colour = "black")). This is an important finishing step in a stripped-back plot because removing too many panel elements can leave the chart without enough visual structure. The explicit axis line restores a frame of reference without reintroducing the full border box.

Two combined approaches are worth knowing. The first uses a single custom theme call:

myplot + theme(
  panel.grid.major = element_blank(),
  panel.grid.minor = element_blank(),
  panel.background = element_blank(),
  axis.line = element_line(colour = "black")
)

The second starts from theme_bw() and then removes the border and grids whilst adding axis lines:

myplot + theme_bw() + theme(
  panel.border = element_blank(),
  panel.grid.major = element_blank(),
  panel.grid.minor = element_blank(),
  axis.line = element_line(colour = "black")
)

Both approaches produce a cleaner chart, though they begin from slightly different defaults. The practical lesson is that {ggplot2} styling is modular, so there is often more than one route to a similar visual result.

This matters because chart design is rarely only about appearance. Cleaner formatting can make a chart easier to read by reducing distractions and placing more emphasis on the data itself. A centred title, a restrained background and the selective use of borders all influence how quickly the eye settles on what is important.

Building Reusable Custom Plot Functions

A third area extends these ideas further by showing how to build custom {ggplot2} functions in R, a topic covered in depth by Sharon Machlis in a tutorial published on Infoworld. The central problem discussed is the mismatch that used to make this awkward: tidyverse functions typically use unquoted column names, whilst base R functions generally expect quoted names. This tension became especially noticeable when users wanted to write their own plotting functions that accepted a data frame and column names as arguments.

The example in that article uses Zillow data containing estimated median home values. After loading {dplyr} and {ggplot2}, a horizontal bar chart is created to show home values by neighbourhood in Boston, with bars ordered from highest to lowest values, outlined in black and filled in blue:

ggplot(data = bos_values, aes(x = reorder(RegionName, Zhvi), y = Zhvi)) +
  geom_col(color = "black", fill = "#0072B2") +
  xlab("") + ylab("") +
  ggtitle("Zillow Home Value Index by Boston Neighborhood") +
  theme_classic() +
  theme(plot.title = element_text(size = 24)) +
  coord_flip()

The next step is to turn that pattern into a function. An initial attempt passes unquoted column names but does not work as intended because of the underlying tension between standard R evaluation and the non-standard evaluation of {ggplot2}. The solution came with the introduction of the tidy evaluation {{ operator, commonly known as "curly-curly", in {rlang} version 0.4.0. As noted in the official tidyverse announcement, this operator abstracts the previous two-step quote-and-unquote process into a single interpolation step. Once library(rlang) is loaded, column references inside the plotting code are wrapped in double curly braces:

library(rlang)
mybarplot <- function(mydf, myxcol, myycol, mytitle) {
  ggplot2::ggplot(data = mydf, aes(x = reorder({{ myxcol }}, {{ myycol }}), y = {{ myycol }})) +
    geom_col(color = "black", fill = "#0072B2") +
    xlab("") + ylab("") +
    coord_flip() +
    ggtitle(mytitle) +
    theme_classic() +
    theme(plot.title = element_text(size = 24))
}

With that change in place, the function can be called with unquoted column names, just as they would appear in many tidyverse functions:

mybarplot(bos_values, RegionName, Zhvi, "Zillow Home Value Index by Boston Neighborhood")

That final point is particularly useful in practice. The resulting plot object can be stored and extended further, for example by adding data labels on the bars with geom_text() and the scales::comma() function. A custom plotting function does not lock the user into a fixed result; it provides a well-designed starting point that can still be extended with additional {ggplot} layers.

Putting the Three Techniques Together in {ggplot2}

Seen as a progression, these examples build on one another in a logical way. The first shows how to centre a title with theme(plot.title = element_text(hjust = 0.5)). The second shows how to simplify a chart by removing grids, borders and background elements whilst restoring axis lines where needed. The third scales those preferences up by packaging them inside a reusable function. What begins as a one-off styling adjustment can therefore become part of a repeatable workflow.

These techniques also reflect a wider culture around R graphics. Resources such as the R Graph Gallery, created by Yan Holtz, have helped make this style of incremental learning more accessible by offering reproducible examples across a wide range of chart types. The gallery presents over 400 R-based graphics, with a strong emphasis on {ggplot2} and the tidyverse, and organises them into nearly 50 chart families and use cases. Its broader message is that effective visualisation is often the result of small, deliberate decisions rather than dramatic reinvention.

For anyone working with {ggplot2}, that is a helpful principle to keep in mind. A centred title may seem minor, just as removing a panel grid may seem cosmetic, yet these changes can improve clarity and consistency across a body of work. When those preferences are wrapped into a function, they also save time and reduce repetition, connecting plot styling directly to good code design.