基于R语言绘制多患者用药时间线图谱的技术求助

在绘制多患者用药时间线时，当p_start/p_end与mid_point_dates坐标重叠时，会出现点被覆盖的情况，导致部分关键时间点无法清晰展示。以下是两种可行的解决方案：

方案1：调整图层顺序与点样式 #

通过控制图层绘制顺序，将需要突出的中间时间点放在最上层，同时给p_start/p_end使用空心带边框的形状，即使重叠也能透过空心看到下方内容。

修改后的代码：

# Load required libraries
library(ggplot2)
library(lubridate)

# Define the data
id = c(214441, 214441, 214441, 214444, 214444, 214446)
med_name = c("morphine", "codeine", "diamorphine", "codeine", "morphine", "drug_X")
p_start = ymd_hms("2010-04-29 12:31:58", "2010-04-29 12:31:58", "2010-04-29 12:31:58", "2010-04-25 12:31:58", "2010-04-25 12:31:58", "2010-04-25 12:31:58")
p_end = ymd_hms("2010-05-19T14:05:00Z", "2010-05-19T14:05:00Z", "2010-05-19 12:31:58", "2010-05-27 12:31:58", "2010-05-27 12:31:58", "2010-05-20 12:31:58")
mid_point_dates = ymd_hms(c("2010-05-09T14:05:00Z", "2010-04-29T14:05:00Z", "2010-05-01T12:52:14Z", "2010-05-13T14:04:00Z", "2010-05-03T14:04:00Z", "2010-04-30T10:34:27Z"))

# Create a data frame
df <- data.frame(id, med_name, p_start, p_end, mid_point_dates)

# Define the y-axis categories
id_categories <- unique(df$id)

ggplot(df, aes(y = id)) +
  # 先绘制用药时间段线段
  geom_segment(aes(x = p_start, xend = p_end, yend = id), size = 1.5, color = "grey80") +
  # 绘制p_start（空心圆形）和p_end（空心方形）
  geom_point(aes(x = p_start), size = 4, shape = 21, fill = "white", color = "black") +
  geom_point(aes(x = p_end), size = 4, shape = 22, fill = "white", color = "black") +
  # 最后绘制中间时间点（实心，覆盖在上方）
  geom_point(aes(x = mid_point_dates, color = med_name, shape = med_name), size = 3) +
  # 样式配置
  scale_color_manual(values = c("purple", "green", "orange", "blue"), name = "Medication") +
  scale_shape_manual(values = c(17, 15, 16, 2), name = "Medication") +
  scale_y_discrete(limits = id_categories) +
  labs(title = "Medication Timeline", x = "Date", y = "Patient ID") +
  theme_minimal()

方案2：重构数据为长格式（推荐） #

将所有时间点（start、mid、end）整合到同一列并标记类型，统一管理图层和样式，后续扩展更灵活。

代码示例：

# Load required libraries
library(ggplot2)
library(lubridate)
library(tidyr)

# Define the data
id = c(214441, 214441, 214441, 214444, 214444, 214446)
med_name = c("morphine", "codeine", "diamorphine", "codeine", "morphine", "drug_X")
p_start = ymd_hms("2010-04-29 12:31:58", "2010-04-29 12:31:58", "2010-04-29 12:31:58", "2010-04-25 12:31:58", "2010-04-25 12:31:58", "2010-04-25 12:31:58")
p_end = ymd_hms("2010-05-19T14:05:00Z", "2010-05-19T14:05:00Z", "2010-05-19 12:31:58", "2010-05-27 12:31:58", "2010-05-27 12:31:58", "2010-05-20 12:31:58")
mid_point_dates = ymd_hms(c("2010-05-09T14:05:00Z", "2010-04-29T14:05:00Z", "2010-05-01T12:52:14Z", "2010-05-13T14:04:00Z", "2010-05-03T14:04:00Z", "2010-04-30T10:34:27Z"))

# 转换为长格式数据
df <- data.frame(id, med_name, p_start, p_end, mid_point_dates) %>%
  pivot_longer(cols = c(p_start, p_end, mid_point_dates),
               names_to = "point_type",
               values_to = "date")

# 定义y轴类别
id_categories <- unique(df$id)

ggplot(df, aes(y = id)) +
  # 绘制用药时间段线段（从去重后的宽格式数据取）
  geom_segment(data = distinct(df, id, med_name, p_start, p_end),
               aes(x = p_start, xend = p_end, yend = id),
               size = 1.5, color = "grey80") +
  # 先绘制start/end点
  geom_point(data = filter(df, point_type %in% c("p_start", "p_end")),
             aes(x = date, shape = point_type),
             size = 4, fill = "white", color = "black") +
  # 最后绘制中间时间点
  geom_point(data = filter(df, point_type == "mid_point_dates"),
             aes(x = date, color = med_name, shape = med_name),
             size = 3) +
  # 样式配置
  scale_color_manual(values = c("purple", "green", "orange", "blue"), name = "Medication") +
  scale_shape_manual(values = c(p_start=21, p_end=22, morphine=17, codeine=15, diamorphine=16, drug_X=2),
                     name = "Point Type") +
  scale_y_discrete(limits = id_categories) +
  labs(title = "Medication Timeline", x = "Date", y = "Patient ID") +
  theme_minimal()

关键说明 #

• 图层顺序：ggplot2中后添加的图层会绘制在上方，将需要突出的中间时间点放在最后，就能覆盖在start/end点之上。
• 形状选择：带填充的空心形状（如shape=21/22）可保留底层视觉信息，实心形状（如15/16/17）用于突出中间时间点。
• 长格式数据：统一管理所有时间点，减少重复代码，后续新增点类型时更易扩展。

内容的提问来源于stack exchange，提问作者db2020