Python新手求助：基于Pivot Table结果绘制分组柱状图

Hey there! I see you've already nailed the pivot table setup and no-show proportion calculations—great work! Let's get that grouped bar chart sorted out, where each health category has separate colored bars for male and female.

Using Seaborn (Recommended for Newbies) #

Seaborn makes this kind of grouped chart super straightforward with its barplot function, which handles gender-based grouping automatically via the hue parameter. Here's how to implement it:

First, make sure you have the required libraries imported (if you haven't already):

import seaborn as sns
import matplotlib.pyplot as plt

Then run this code to generate your chart:

# Set a clean, readable plot style (optional but improves aesthetics)
sns.set_style("whitegrid")

# Create the grouped bar chart
plt.figure(figsize=(10, 6))  # Adjust plot size to your preference
ax = sns.barplot(
    x='category', 
    y='no_show_prop', 
    hue='gender', 
    data=data_pv,
    palette='Set2'  # Pick a color palette you like
)

# Add clear labels and a title
ax.set_title('No-Show Proportion by Health Category and Gender', fontsize=14)
ax.set_xlabel('Health Condition Category', fontsize=12)
ax.set_ylabel('No-Show Proportion (%)', fontsize=12)

# Adjust legend position to avoid overlapping bars
plt.legend(title='Gender', bbox_to_anchor=(1.05, 1), loc='upper left')

# Ensure labels don't get cut off
plt.tight_layout()
# Display the plot
plt.show()

This will give you exactly what you want: each category on the X-axis, with two distinct bars (one for F, one for M) showing the no-show percentage. The hue parameter is the key here—it tells Seaborn to split bars by the gender column automatically.

If You Prefer Matplotlib (For More Low-Level Control) #

If you want to understand the underlying mechanics, you can build the chart manually with Matplotlib:

import matplotlib.pyplot as plt
import numpy as np

# Get unique categories and genders from your data
categories = data_pv['category'].unique()
genders = data_pv['gender'].unique()
x = np.arange(len(categories))  # Base X positions for categories
bar_width = 0.35  # Width of each individual bar

# Initialize the plot
fig, ax = plt.subplots(figsize=(10, 6))

# Loop through genders to plot their bars
for idx, gender in enumerate(genders):
    # Filter data for the current gender
    gender_subset = data_pv[data_pv['gender'] == gender]
    # Plot bars with an offset to avoid overlapping
    ax.bar(x + idx*bar_width, gender_subset['no_show_prop'], bar_width, label=gender)

# Customize the plot
ax.set_title('No-Show Proportion by Health Category and Gender', fontsize=14)
ax.set_xlabel('Health Condition Category', fontsize=12)
ax.set_ylabel('No-Show Proportion (%)', fontsize=12)
# Set X-ticks to sit between grouped bars
ax.set_xticks(x + bar_width/2)
ax.set_xticklabels(categories)
ax.legend(title='Gender')

plt.tight_layout()
plt.show()

Why Your Groupby Attempt Might Have Failed #

Chances are your groupby approach didn't work because you were trying to plot directly from the grouped object without reshaping it into a format that's easy to plot. The Seaborn method avoids this headache entirely, as it works seamlessly with your existing data_pv DataFrame (which is already in the perfect "long format" for grouped plots).

Give either method a try—they should get you the chart you're aiming for!

内容的提问来源于stack exchange，提问作者LOR121