Adding Black Borders to Data-Filled Points in ggplot2 Scatterplots: Core Techniques and Implementation

Introduction and Problem Context

In data visualization, scatterplots are a common tool for displaying relationships between two variables. When using the ggplot2 package in R, users often need to fill points with colors based on data and add borders to enhance visual contrast. However, beginners frequently encounter issues such as failed fill color mapping or abnormal legend display when attempting this effect. This article, based on a typical Q&A from Stack Overflow, delves into how to correctly add black borders to data-filled points in ggplot2 and avoid extra legend entries.

Core Knowledge: Shape Parameters and Color Separation

In ggplot2, point appearance is controlled by the shape parameter, with different shapes corresponding to different fill and border behaviors. For example, shapes 21 to 25 allow independent setting of fill color (fill) and border color (colour), while shapes 1 to 20 use a single color. This is key to achieving data-filled points with borders. The following code demonstrates the correct approach:

library(ggplot2)
df <- data.frame(id = runif(12), x = 1:12, y = runif(12))
ggplot(df, aes(x = x, y = y)) +
  geom_point(aes(fill = id), colour = "black", shape = 21, size = 5)

Here, aes(fill = id) maps the fill color to the id variable, while colour = "black" sets the border color as an absolute black outside the aesthetic mapping, preventing legend generation. Using shape = 21 (equivalent to pch = 21) enables separation of fill and border. In contrast, an incorrect method like geom_point(aes(fill = id, colour = "black"), size = 12) maps colour to a constant string, causing fill color to fail because ggplot2 defaults some shapes to not support the fill aesthetic.

Legend Control and Advanced Adjustments

By default, ggplot2 may use a colorbar instead of a point-based legend. To display a traditional point legend, the guides function can be used:

g0 <- ggplot(df, aes(x = x, y = y)) +
  geom_point(aes(fill = id), colour = "black", shape = 21, size = 5)
g0 + guides(fill = "legend")

Further, scale_fill_continuous allows customization of legend breaks and labels:

g0 + scale_fill_continuous(guide = "legend", breaks = seq(0.2, 0.8, by = 0.1))

This ensures the legend accurately reflects the data range, enhancing visualization professionalism.

Supplementary Method: Layering Technique

Referencing other answers, an alternative approach involves using two geom_point layers: one for filled points and another for borders. Example code is as follows:

ggplot(df, aes(x = x, y = y)) +
  geom_point(aes(colour = id), size = 12) +
  geom_point(shape = 1, size = 12, colour = "black")

This method adds borders via shape = 1 (hollow circle), but may introduce extra legend entries that need manual suppression. Although flexible, it increases code complexity and is less intuitive than the shape parameter method.

In-Depth Analysis: Difference Between Aesthetic Mapping and Absolute Setting

Understanding the distinction between setting colors inside and outside aes() is crucial. Inside aes(), e.g., aes(colour = "black"), color is mapped as a variable, leading to legend generation; outside aes(), e.g., colour = "black", color is set as an absolute constant, not affecting the legend. This explains why the best answer places colour outside aes() to achieve a border without a legend.

Practical Applications and Best Practices

In real-world projects, the shape parameter method is recommended due to its concise code, high performance, and ease of maintenance. Ensure to select shapes that support fill (e.g., 21-25) and set border color outside aes(). For complex legends, fine-tune using guides and scale_* functions. Avoid common pitfalls, such as using shapes that do not support fill or misplacing aesthetic mappings.

Conclusion

Through this article, we have uncovered the core mechanism for adding black borders to data-filled points in ggplot2: leveraging specific shape parameters to separate fill and border, and correctly setting aesthetic mappings. Best practices include using shape = 21, defining border color outside aes(), and controlling legends via guides. These techniques not only solve the specific problem but also deepen understanding of ggplot2 visualization principles, contributing to more efficient data analysis workflows.