如何在R中使用gganimate制作三元图动画？遇实现障碍求助

核心问题分析 #

1. 三元数据不合法：三元图要求x + y + z = 1，你用runif()生成的三个变量和不为1，导致stat_density_tern出现非有限值警告，同时干扰动画渲染逻辑。
2. transition_states参数顺序错误：该函数的正确语法为transition_states(states, transition_length, state_length)，你将transition_length放在了states参数前，导致函数解析失败。
3. 统计层分组缺失：stat_density_tern需要明确按time分组，否则无法为每个时间点单独计算密度分布。

修正后的代码示例 #

1. 先实现基础动态点效果

library(tidyverse)
library(gganimate)
library(ggtern)

# 修正三元数据：确保x+y+z=1
data <- tibble(x = runif(200), y = runif(200), z = runif(200), time = rep(1:10, 20)) %>%
  rowwise() %>%
  mutate(total = x + y + z,
         x = x/total, y = y/total, z = z/total) %>%
  ungroup()

# 动态点动画
ggtern(data, aes(x, y, z)) +
  geom_point() +
  theme_rgbw() +
  # 修正参数顺序：先指定states，再设置过渡时长
  transition_states(states = time, transition_length = 1, state_length = 1) +
  labs(title = "Time: {closest_state}")

2. 带密度层的完整动画

ggtern(data, aes(x, y, z, group = time)) +
  geom_point() +
  theme_rgbw() +
  stat_density_tern(geom = 'polygon',
                    aes(fill = ..level.., alpha = ..level..),
                    bdl = 0.001,
                    # 强制按time分组计算每个时间点的密度
                    group = time) +
  scale_fill_gradient(low = "blue", high = "red") +
  guides(color = "none", fill = "none", alpha = "none") +
  transition_states(states = time, transition_length = 1, state_length = 1) +
  labs(title = "Time: {closest_state}")

3. 连续时间过渡版本（用transition_time）

如果希望时间过渡更平滑，可替换为transition_time：

ggtern(data, aes(x, y, z)) +
  geom_point() +
  theme_rgbw() +
  stat_density_tern(geom = 'polygon',
                    aes(fill = ..level.., alpha = ..level.., group = time),
                    bdl = 0.001) +
  scale_fill_gradient(low = "blue", high = "red") +
  guides(color = "none", fill = "none", alpha = "none") +
  transition_time(time) +
  labs(title = "Time: {frame_time}")

内容的提问来源于stack exchange，提问作者LDT