R语言时间序列分析：如何让柱状图Y轴显示月度投诉总数

Looks like you're trying to plot monthly complaint volumes but ended up with complaint names on the Y-axis—let's fix that! The core issue here is how you're mapping variables to the ggplot aesthetics and using statistical functions.

What was wrong with your original code? #

• You mapped Customer.Complaint (a string variable with complaint text) to the Y-axis. ggplot treats this as a discrete variable, so it displays individual complaint names instead of counts.
• Using stat_summary(fun.y=sum) on a string variable doesn't make sense—we need to count complaints per month, not sum text values.
• The scale_y_continuous(labels = fun.y= length) line has incorrect syntax and doesn't align with your goal of showing counts.

Method 1: Pre-aggregate data first (recommended for clarity) #

First, calculate monthly complaint counts using dplyr, then plot the aggregated data. This makes your data transformations explicit and easier to debug.

library(ggplot2)
library(scales)
library(dplyr)

# Convert date to proper format
comcast$Date <- gsub('-', '/', comcast$Date)
comcast$Date <- as.Date(comcast$Date, '%d/%m/%Y')

# Create monthly date grouping and count complaints per month
comcast_monthly <- comcast %>%
  mutate(Date_by_month = as.Date(cut(Date, breaks = 'month'))) %>%
  group_by(Date_by_month) %>%
  summarise(Complaint_Count = n()) # Count number of complaints per group

# Plot the trend
ggplot(comcast_monthly, aes(x = Date_by_month, y = Complaint_Count)) +
  geom_bar(stat = 'identity', fill = '#2c3e50') +
  scale_x_date(labels = date_format("%Y-%m"), breaks = '1 month') +
  labs(title = "Monthly Comcast Complaint Trend",
       x = "Month",
       y = "Number of Complaints") +
  theme_minimal()

Method 2: Use stat_count directly in ggplot (no pre-aggregation) #

If you want to handle counting entirely within the ggplot call, use stat_count—it automatically counts the number of observations (complaints) for each month. You don't need to map a Y variable here.

library(ggplot2)
library(scales)

# Date conversion steps
comcast$Date <- gsub('-', '/', comcast$Date)
comcast$Date <- as.Date(comcast$Date, '%d/%m/%Y')
comcast$Date_by_month <- as.Date(cut(comcast$Date, breaks='month'))

# Plot with automatic counting
ggplot(comcast, aes(x = Date_by_month)) +
  stat_count(geom = 'bar', fill = '#2c3e50') +
  scale_x_date(labels = date_format("%Y-%m"), breaks = '1 month') +
  labs(title = "Monthly Comcast Complaint Trend",
       x = "Month",
       y = "Number of Complaints") +
  theme_minimal()

Key Takeaways #

• Map counts to Y-axis: Either pre-calculate the monthly counts (Method 1) or let stat_count handle it (Method 2)—don't map raw complaint text to Y.
• Use the right statistical function: n() (count) is what you need here, not sum() (which is for numerical values).

内容的提问来源于stack exchange，提问作者Darkstar Dream