如何使用ggplot绘制平滑填充图？基于现有数据集与代码求优化方案

Hey there! Let's spruce up that plot into a smoother, more polished filled visualization. Your current code creates a stacked proportional bar chart—here's how to shift to a sleek, smooth look, plus extra tweaks to make it pop:

Step 1: Prep your data for smooth stacked areas #

First, we need to calculate proportional values (matching your position="fill" behavior) and their cumulative sums, which is essential for creating stacked smooth areas:

library(dplyr)
library(ggplot2)

# Calculate proportional and cumulative proportional values per period
df_smooth <- df_test %>%
  group_by(period) %>%
  mutate(prop = value / sum(value),  # Get percentage of total value per emotion
         cum_prop = cumsum(prop)) %>%  # Cumulative sum for proper stacking
  ungroup() %>%
  arrange(period, emotie)  # Sort to ensure stacking order is correct

Step 2: Create your smooth filled plot #

We'll replace rigid bars with smooth, continuous fills. Below are two options depending on whether you want to keep period as your x-axis or shift to a continuous emotion axis:

Option 1: Faceted smooth stacked area plot (continuous emotion axis)

This view shows the flow of value proportions across emotion categories for each period, with smooth curves instead of sharp bar edges:

ggplot(df_smooth, aes(x = emotie, y = prop, fill = factor(emotie))) +
  geom_area(position = "stack", stat = "smooth", method = "loess", alpha = 0.8) +
  facet_wrap(~period) +  # Split into separate panels for each period
  labs(x = "Emotion", y = "Proportion of Total Value", fill = "Emotion") +
  theme_minimal() +
  scale_fill_viridis_d(option = "plasma")  # Colorblind-friendly palette

Option 2: Smooth "bar-like" filled shapes (period as x-axis)

If you want to retain period as your x-axis (matching your original plot structure) but with rounded, smooth edges:

ggplot(df_smooth, aes(x = as.numeric(period), y = cum_prop, group = factor(emotie), fill = factor(emotie))) +
  geom_ribbon(aes(ymin = lag(cum_prop, default = 0), ymax = cum_prop), alpha = 0.8, stat = "smooth", method = "loess") +
  scale_x_continuous(breaks = c(1,2), labels = c("Period 1", "Period 2")) +
  labs(x = "Period", y = "Cumulative Proportion", fill = "Emotion") +
  theme_minimal() +
  scale_fill_viridis_d(option = "mako")

Extra visual polish tips #

• Adjust the alpha parameter (e.g., alpha = 0.7) to make fill colors semi-transparent, so overlapping smooth areas are easier to distinguish
• Use a colorblind-friendly palette like scale_fill_viridis_d() or scale_fill_brewer(palette = "Set2") instead of the default colors
• Add descriptive axis labels and a clear legend title to help viewers understand the plot at a glance
• Swap theme_minimal() for theme_classic() or theme_light() if you prefer a cleaner, less grid-heavy aesthetic
• For the faceted plot, add theme(strip.text = element_text(size = 12, face = "bold")) to make period labels stand out

内容的提问来源于stack exchange，提问作者Henk Straten