geom_ribbon绘制年龄占比：空白区域与额外凸起修复

初始问题 #

我想要展示不同年龄组的占比随时间的变化趋势。目前已获取各年龄组的个体占比（pAGE），以及用于geom_ribbon绘图的各年龄组累积占比区间值（cpAgeMin和cpAgeMax）。尽管所有年龄组占比总和为1，但绘图中部分年份存在无数据的空白区域，这并非数据问题，而是ggplot的绘图问题。需要修复此问题，确保每年0-1区间都被某一年龄组完全覆盖。

示例数据及绘图代码 #

# 数据子集
df <- structure(list(site = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
                                    1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
                                    1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L), 
                                  levels = "G", class = "factor"), 
                 year = c(1988, 1988, 1989, 1989, 1989, 1990, 1990, 1991, 1991, 1991, 
                          1992, 1992, 1992, 1993, 1993, 1993, 1998, 1998, 1998, 1999, 
                          1999, 1999, 2000, 2000, 2000, 2001, 2001, 2001, 2002, 2002, 
                          2003, 2003, 2003), 
                 age = c(`9104` = 1, `9341` = 2, `9292` = 1, `9632` = 2, `9960` = 4, 
                         `9543` = 1, `9857` = 3, `9685` = 1, `9968` = 2, `10169` = 3, 
                         `9858` = 1, `10127` = 2, `10212` = 3, `10009` = 1, `10284` = 2, 
                         `10522` = 6, `10464` = 1, `10605` = 2, `10830` = 3, `10556` = 1, 
                         `10744` = 2, `11056` = 3, `10687` = 1, `11061` = 2, `11279` = 5, 
                         `10912` = 1, `11109` = 2, `11197` = 3, `11065` = 1, `11194` = 2, 
                         `11154` = 1, `11255` = 2, `11324` = 3), 
                 pAGE = c(0.972, 0.028, 0.964, 0.005, 0.031, 0.823, 0.177, 0.921, 0.074, 
                          0.004, 0.846, 0.045, 0.109, 0.833, 0.155, 0.012, 0.927, 0.054, 
                          0.019, 0.784, 0.21, 0.005, 0.958, 0.017, 0.025, 0.852, 0.124, 
                          0.024, 0.913, 0.087, 0.909, 0.073, 0.019), 
                 cpAgeMin = c(0, 0.972, 0, 0.964, 0.969, 0, 0.823, 0, 0.921, 0.996, 0, 
                              0.846, 0.891, 0, 0.833, 0.988, 0, 0.927, 0.981, 0, 0.784, 
                              0.995, 0, 0.958, 0.975, 0, 0.852, 0.976, 0, 0.913, 0, 0.909, 
                              0.981), 
                 cpAgeMax = c(0.972, 1, 0.964, 0.969, 1, 0.823, 1, 0.921, 0.996, 1, 0.846, 
                              0.891, 1, 0.833, 0.988, 1, 0.927, 0.981, 1, 0.784, 0.995, 1, 
                              0.958, 0.975, 1, 0.852, 0.976, 1, 0.913, 1, 0.909, 0.981, 1),
                 seg = c(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 
                         2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2)), 
            row.names = c(NA, -33L), class = c("tbl_df", "tbl", "data.frame"))

# 绘图代码
ggplot(data = df) + 
  geom_ribbon(aes(x = year, ymin = cpAgeMin, ymax = cpAgeMax, group = interaction(seg, factor(age)), fill = factor(age), col = factor(age))) +
  theme_bw()

初始绘图效果 #

更新问题 #

在为第二个站点应用@Andy Baxter的解决方案后，绘图中数据范围外出现了额外的“凸起”，如何消除这些凸起？

更新后绘图问题 #

解决初始空白问题 #

空白区域出现的原因是geom_ribbon按seg+age分组时，会将不连续年份的同分组数据连接，导致跨年份填充断裂。正确做法是按年份+seg+年龄组分组，确保每个年份内的年龄组区间独立连续：

修改绘图代码的group参数：

ggplot(data = df) + 
  geom_ribbon(aes(x = year, ymin = cpAgeMin, ymax = cpAgeMax, 
                  group = interaction(year, seg, factor(age)), 
                  fill = factor(age), col = factor(age))) +
  theme_bw()

这样每个年份的每个年龄组区间都会被单独处理，确保0-1区间完全覆盖，不会出现空白。

解决额外凸起问题 #

数据范围外的凸起是因为geom_ribbon默认会延伸连接分组的首尾点，导致超出数据年份的区域出现异常填充。推荐以下两种解决方式：

方法1：限制x轴范围到数据实际年份 #

使用coord_cartesian()裁剪显示范围，仅展示数据存在的年份区间，不会过滤原始数据：

ggplot(data = df) + 
  geom_ribbon(aes(x = year, ymin = cpAgeMin, ymax = cpAgeMax, 
                  group = interaction(year, seg, factor(age)), 
                  fill = factor(age), col = factor(age))) +
  coord_cartesian(xlim = range(df$year)) +
  theme_bw()

方法2：为分组添加首尾闭合点 #

若需保留x轴扩展范围，可在数据中为每个seg+age分组添加首尾年份的边界点（将cpAgeMin和cpAgeMax设为与相邻年份一致或NA），但此方法操作繁琐，优先推荐方法1。

内容的提问来源于stack exchange，提问作者tnt