Implementing Dual Y-Axis Visualizations in ggplot2: Methods and Best Practices

Introduction

In data visualization practice, there is often a need to simultaneously display two data series with different dimensions and numerical ranges. ggplot2, as the most popular visualization package in R, has provided the sec_axis() function since version 2.2.0 to implement dual Y-axis functionality. However, this visualization approach remains controversial in academic circles, requiring a balance between technical implementation and visualization principles.

Basic Implementation Principles of Dual Y-Axes

The dual Y-axes in ggplot2 are not truly independent coordinate axes, but rather secondary axes generated through mathematical transformations based on the primary axis. This design ensures clear mathematical relationships between the axes, avoiding potential misinterpretation from arbitrary scaling.

Basic implementation code:

library(ggplot2)

# Create sample data
data <- data.frame(
  x = 1:10,
  primary = runif(10, 0, 100),
  secondary = runif(10, 0, 10)
)

# Calculate scaling factor
scale_factor <- max(data$primary) / max(data$secondary)

# Build dual Y-axis chart
ggplot(data, aes(x = x)) +
  geom_line(aes(y = primary), color = "blue") +
  geom_line(aes(y = secondary * scale_factor), color = "red") +
  scale_y_continuous(
    name = "Primary Variable",
    sec.axis = sec_axis(~ . / scale_factor, name = "Secondary Variable")
  )

Practical Application Case Analysis

Consider a common business analysis scenario: the need to simultaneously display product sales (bar chart) and conversion rates (line chart). Since sales and conversion rates have vastly different numerical ranges, direct overlay would make one series almost invisible.

The core of the solution lies in correctly calculating the scaling factor:

# Sample data
sales_data <- data.frame(
  month = month.abb[1:6],
  sales = c(1000, 1200, 800, 1500, 2000, 1800),
  conversion_rate = c(0.15, 0.18, 0.12, 0.20, 0.25, 0.22)
)

# Calculate appropriate scaling factor
conversion_scale <- max(sales_data$sales) / max(sales_data$conversion_rate)

ggplot(sales_data, aes(x = month)) +
  geom_col(aes(y = sales), fill = "steelblue", alpha = 0.7) +
  geom_line(aes(y = conversion_rate * conversion_scale), 
            color = "red", size = 1.5, group = 1) +
  geom_point(aes(y = conversion_rate * conversion_scale), 
             color = "red", size = 3) +
  scale_y_continuous(
    name = "Sales (Units)",
    sec.axis = sec_axis(~ . / conversion_scale, 
                       name = "Conversion Rate (%)",
                       labels = scales::percent)
  ) +
  theme_minimal() +
  labs(title = "Monthly Sales and Conversion Rate Trends")

Axis Style Customization

To enhance chart readability, differential styling can be applied to the two axes using the theme() function:

ggplot(sales_data, aes(x = month)) +
  geom_col(aes(y = sales), fill = "#3498db", alpha = 0.8) +
  geom_line(aes(y = conversion_rate * conversion_scale), 
            color = "#e74c3c", size = 1.2, group = 1) +
  scale_y_continuous(
    name = "Sales",
    sec.axis = sec_axis(~ . / conversion_scale, name = "Conversion Rate")
  ) +
  theme(
    axis.title.y.left = element_text(color = "#3498db", size = 12),
    axis.text.y.left = element_text(color = "#3498db"),
    axis.title.y.right = element_text(color = "#e74c3c", size = 12),
    axis.text.y.right = element_text(color = "#e74c3c")
  )

Alternative Approach: Faceted Plots

When dual Y-axes might cause misinterpretation, faceted plotting provides a safer alternative. This method plots the two data series in separate subplots, sharing the X-axis but having independent Y-axes.

library(tidyr)

# Reshape data to long format
long_data <- sales_data %>%
  pivot_longer(cols = c(sales, conversion_rate), 
               names_to = "metric", 
               values_to = "value")

# Create faceted chart
ggplot(long_data, aes(x = month, y = value)) +
  geom_col(data = filter(long_data, metric == "sales"), 
           fill = "steelblue") +
  geom_line(data = filter(long_data, metric == "conversion_rate"), 
            color = "red", size = 1.2, group = 1) +
  facet_grid(metric ~ ., scales = "free_y") +
  theme_minimal()

Best Practices and Considerations

When using dual Y-axis visualizations, special attention should be paid to the following aspects:

Mathematical Relationship Clarity: Ensure clear mathematical transformation relationships exist between the two axes, avoiding arbitrary scaling.

Visual Guidance: Use colors, legends, and axis styles to clearly distinguish between the two data series, reducing cognitive load for readers.

Data Range Matching: Appropriately select scaling factors so both data series have suitable display ranges in the chart.

Contextual Explanation: Explain the relationship between the two variables and the scaling principle in chart titles or annotations.

Conclusion

ggplot2's sec_axis() function provides powerful technical support for implementing dual Y-axis visualizations. Through correct mathematical transformations and appropriate visual design, information-rich composite charts can be created while maintaining data accuracy. However, developers should always consider visualization principles and data interpretation accuracy, choosing safer alternatives like faceted plotting when necessary.

In practical applications, it's recommended to first evaluate whether dual Y-axes truly contribute to the data story, rather than using them simply because they are technically feasible. Dual Y-axes deliver maximum value when there are clear causal or comparative relationships between the two variables.