R语言中ggplot图表无法正确显示星期几的问题求助

问题详情 #

• 绘制现象：用ggplot绘制星期几与总步数的柱状图时，x轴的星期几无法区分，呈现混乱的一团
• 数据状态：day_of_week为有序因子（Ord.factor），包含7个层级，但存在大量NA值
• 尝试过的调整代码：

daily_activity_cleaned$day_of_week <- wday(x, week_start ="lubridate.week.start", 7)

• 当前绘图代码：

ggplot(data = daily_activity_cleaned) + aes(x = day_of_week, y = total_steps) + 
    geom_col(fill = 'blue') + 
    labs(x = 'Day of week', y = 'Total Steps', title = 'Total steps taken in a week')

• 数据类型输出：

str(daily_activity_cleaned$day_of_week)
# Ord.factor w/ 7 levels "Sun"<"Mon"<"Tue"<..: NA NA NA NA NA NA NA NA NA NA ...

问题根源 #

1. 大量NA值会在绘图时生成无效的x轴元素，干扰正常的星期分组显示
2. 使用wday()时的参数设置有误，没有正确生成带名称的有序因子，或未校准层级顺序

解决步骤 #

1. 清理无效的NA数据 #

先过滤掉day_of_week为NA的行，避免无效数据干扰：

daily_activity_cleaned <- daily_activity_cleaned[!is.na(daily_activity_cleaned$day_of_week), ]

2. 重新生成正确的星期有序因子 #

假设你的x是日期列，用lubridate生成带星期名称的有序因子，明确指定层级顺序：

library(lubridate)
# 生成英文星期名称，指定周日为一周起始
daily_activity_cleaned$day_of_week <- factor(
  wday(x, label = TRUE, week_start = 7),
  levels = c("Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat"),
  ordered = TRUE
)

# 如果需要中文星期，用以下代码
daily_activity_cleaned$day_of_week <- factor(
  wday(x, label = TRUE, week_start = 7, locale = "zh_CN.UTF-8"),
  levels = c("周日", "周一", "周二", "周三", "周四", "周五", "周六"),
  ordered = TRUE
)

3. 优化绘图逻辑（避免堆叠混乱） #

如果原数据是按日期记录的单日步数，直接用geom_col会堆叠同一星期的多条数据，导致显示异常。建议先按星期汇总总步数再绘图：

# 方法1：用dplyr先聚合数据
library(dplyr)
weekly_summary <- daily_activity_cleaned %>%
  group_by(day_of_week) %>%
  summarise(total_steps = sum(total_steps, na.rm = TRUE))

ggplot(weekly_summary) +
  aes(x = day_of_week, y = total_steps) +
  geom_col(fill = "blue") +
  labs(x = "星期", y = "总步数", title = "一周总步数统计")

# 方法2：直接在ggplot中用stat_summary汇总
ggplot(daily_activity_cleaned) +
  aes(x = day_of_week, y = total_steps) +
  stat_summary(fun = sum, geom = "col", fill = "blue") +
  labs(x = "星期", y = "总步数", title = "一周总步数统计")

4. 验证因子层级 #

执行以下代码确认星期的排序符合预期：

levels(daily_activity_cleaned$day_of_week)

内容的提问来源于stack exchange，提问作者rayban0622