如何在R中绘制随时间变化的相对比例图？含研究绘图需求

Got it, let's break down how to create this exact chart you need for your genre annotation study. I’ve worked through similar visualization tasks for behavioral experiments, so here’s a practical, step-by-step approach with code examples to get you sorted:

First, let's align on the key details to make sure we're building the right thing:

• X-axis: Participants ordered chronologically, with all tags from the same participant sharing the exact same X position (e.g., Tagid 1 and 2 from Participant A both map to X=3 if they were the 3rd participant)
• Y-axis: Frequency of your target tag word (either raw count of how many times the participant used the word, or its proportion relative to all tags they submitted)
• Dataset: 10 songs' genre annotation data, with each participant's tags stored separately

Before plotting, you’ll need to clean and restructure your data into a plot-friendly format:

• Assign a unique chronological participant ID (e.g., Participant_1, Participant_2) to every tag entry, matching the order participants completed the task
• Add a time order index (e.g., 1, 2, 3) for each participant — this will be your X-axis value (same index for all tags from one participant)
• Calculate the target word's frequency per participant (raw count or proportion)

This is a go-to setup for scientific plotting. Below is a complete example with simulated data matching your use case:

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

# Simulate your actual dataset structure
raw_data = {
    "Participant_ID": ["P1", "P1", "P2", "P3", "P3", "P3", "P4"],
    "Tagid": [1, 2, 1, 1, 2, 3, 1],
    "Tag_Word": ["jazz", "blues", "jazz", "rock", "jazz", "jazz", "blues"],
    "Time_Order": [1, 1, 2, 3, 3, 3, 4]  # X-axis position for each participant
}
df = pd.DataFrame(raw_data)

# Define your target tag word
target_tag = "jazz"

# Calculate raw frequency (count of target word per participant)
frequency_df = df[df["Tag_Word"] == target_tag].groupby("Participant_ID")["Tagid"].count().reset_index(name="Target_Count")
# Merge with time order data (ensure one X value per participant)
frequency_df = frequency_df.merge(df[["Participant_ID", "Time_Order"]].drop_duplicates(), on="Participant_ID")

# Plot the chart
plt.figure(figsize=(12, 6))

# Bar chart for overall participant frequency
sns.barplot(x="Time_Order", y="Target_Count", data=frequency_df, color="#2ca02c", alpha=0.7)

# Optional: Add scatter points to show individual tag occurrences (same X for same participant)
individual_tags = df[df["Tag_Word"] == target_tag]
plt.scatter(x=individual_tags["Time_Order"], y=[0.3]*len(individual_tags), color="#ff7f0e", s=50, label="Individual Tag Use")

# Customize labels and styling
plt.xlabel("Participants (Chronological Order)", fontsize=12)
plt.ylabel(f"Frequency of '{target_tag}'", fontsize=12)
plt.title(f"'{target_tag}' Usage Across 10-Song Genre Annotation Participants", fontsize=14)
plt.xticks(rotation=45)
plt.legend()
plt.tight_layout()
plt.show()

If you want Y-axis to show the proportion of the target word relative to all tags a participant used, modify the frequency calculation like this:

# Calculate total tags per participant
total_tags = df.groupby("Participant_ID")["Tagid"].count().reset_index(name="Total_Tags")
# Calculate target word count per participant
target_counts = df[df["Tag_Word"] == target_tag].groupby("Participant_ID")["Tagid"].count().reset_index(name="Target_Count")
# Merge and compute proportion
frequency_df = pd.merge(total_tags, target_counts, on="Participant_ID", how="left").fillna(0)
frequency_df["Target_Proportion"] = frequency_df["Target_Count"] / frequency_df["Total_Tags"]
# Merge with time order data
frequency_df = frequency_df.merge(df[["Participant_ID", "Time_Order"]].drop_duplicates(), on="Participant_ID")

Then use Target_Proportion as your Y-axis value in the plot.

If you prefer R for your analysis, here's an equivalent setup:

library(ggplot2)
library(dplyr)

# Simulate data
raw_data <- data.frame(
  Participant_ID = c("P1", "P1", "P2", "P3", "P3", "P3", "P4"),
  Tagid = c(1, 2, 1, 1, 2, 3, 1),
  Tag_Word = c("jazz", "blues", "jazz", "rock", "jazz", "jazz", "blues"),
  Time_Order = c(1, 1, 2, 3, 3, 3, 4)
)

target_tag <- "jazz"

# Calculate frequency
frequency_df <- raw_data %>%
  filter(Tag_Word == target_tag) %>%
  group_by(Participant_ID) %>%
  summarise(Target_Count = n()) %>%
  left_join(raw_data %>% distinct(Participant_ID, Time_Order), by = "Participant_ID")

# Plot
ggplot(frequency_df, aes(x = factor(Time_Order), y = Target_Count)) +
  geom_bar(stat = "identity", fill = "#2ca02c", alpha = 0.7) +
  geom_jitter(data = raw_data %>% filter(Tag_Word == target_tag), 
              aes(x = factor(Time_Order), y = 0.3), 
              color = "#ff7f0e", size = 3, alpha = 0.7) +
  labs(x = "Participants (Chronological Order)",
       y = paste("Frequency of '", target_tag, "'", sep = ""),
       title = "'jazz' Usage Across 10-Song Genre Annotation Participants") +
  theme(axis.text.x = element_text(rotation = 45, hjust = 1))

内容的提问来源于stack exchange，提问作者jbouwman