Matlab：按试验编号分组子集化并绘制列均值曲线

Hey there! This is a really typical problem when dealing with experimental data that's not perfectly structured, but it's totally solvable with standard data manipulation and visualization tools. Let's walk through how to do this in both R and Python—pick whichever you're more comfortable with!

用R实现（dplyr + ggplot2） #

This combo is perfect for tidy data workflows, which aligns exactly with your long-format dataset.

1. Load required libraries
   First, make sure you've got these packages installed (if not, use install.packages(c("dplyr", "ggplot2"))):

   library(dplyr)
   library(ggplot2)
   

2. Import your data
   Assume your data is in a CSV file—adjust the path and column names as needed (your first column is measurements, second is trial IDs):

   # Replace "your_data.csv" with your actual file path
   df <- read.csv("your_data.csv", header = FALSE, col.names = c("measurement", "trial_id"))
   

3. Calculate group means
   Group by trial ID, compute the mean measurement, and handle any missing values in the measurements with na.rm = TRUE:

   summary_df <- df %>%
     group_by(trial_id) %>%
     summarize(mean_measurement = mean(measurement, na.rm = TRUE))
   

4. Handle discontinuous/missing trials (optional but useful)
   If you want your x-axis to show all trial IDs between the minimum and maximum (including missing ones), convert trial_id to a factor with complete levels:

   # Generate all possible trial IDs in the range
   full_trial_range <- min(df$trial_id):max(df$trial_id)
   summary_df$trial_id <- factor(summary_df$trial_id, levels = full_trial_range)
   

5. Plot the mean curve
   Use ggplot2 to draw the line and points—group = 1 ensures the line connects all points even when trial_id is a factor:

   ggplot(summary_df, aes(x = trial_id, y = mean_measurement, group = 1)) +
     geom_line(color = "#2c3e50", linewidth = 1) +
     geom_point(color = "#e74c3c", size = 3) +
     labs(x = "试验编号", y = "均值测量值", title = "各试验组均值曲线") +
     theme_minimal()
   

用Python实现（pandas + seaborn/matplotlib） #

Pandas makes grouping and aggregation straightforward, and seaborn simplifies creating clean plots.

1. Load required libraries

   import pandas as pd
   import matplotlib.pyplot as plt
   import seaborn as sns
   
   # Set a clean plot theme
   sns.set_theme(style="whitegrid")
   

2. Import your data
   Again, adjust the file path—here we use names to define column names since your data doesn't have headers:

   df = pd.read_csv("your_data.csv", header=None, names=["measurement", "trial_id"])
   

3. Calculate group means
   Group by trial ID and compute the mean. Pandas automatically skips missing values by default, but you can explicitly use skipna=True for clarity:

   summary_df = df.groupby("trial_id")["measurement"].mean().reset_index()
   

4. Handle discontinuous/missing trials (optional)
   To include missing trial IDs on the x-axis (showing gaps where there's no data), create a complete range of trial IDs and merge with your summary data:

   min_trial = df["trial_id"].min()
   max_trial = df["trial_id"].max()
   
   # Create a dataframe with all trial IDs in the range
   full_trials = pd.DataFrame({"trial_id": range(min_trial, max_trial + 1)})
   
   # Merge to keep all trial IDs (missing means will be NaN)
   summary_df = pd.merge(full_trials, summary_df, on="trial_id", how="left")
   

5. Plot the mean curve
   Seaborn will automatically handle gaps where means are NaN (the line will break, which is intuitive for missing trials):

   plt.figure(figsize=(10, 6))
   sns.lineplot(data=summary_df, x="trial_id", y="measurement", marker="o", color="#2c3e50", linewidth=1, markersize=7)
   plt.xlabel("试验编号")
   plt.ylabel("均值测量值")
   plt.title("各试验组均值曲线")
   plt.show()
   

Key Notes #

• Handling missing measurements: Both methods include na.rm/skipna to ignore any missing values within a trial's measurements—adjust this if you want to exclude entire trials with missing data instead.
• Missing trials: The optional steps to include full trial ranges make your plot more informative, as it clearly shows which trials have no data. If you don't mind missing trials being omitted from the x-axis, you can skip those steps.

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