Matplotlib极坐标柱状图：实现中心空心环（甜甜圈孔）

Hey there, let's sort out that donut hole for your polar bar chart! The issue with setting bottom=1 directly is that it shifts your bars upward by 1 (making their total height 1 + original value), which isn't what you want. Instead, we need to keep your data values intact while leaving the 0-1 inner region blank. Here are two straightforward solutions:

Solution 1: Adjust Bar Position & Height (Cleanest for Your Data) #

Since your data doesn't include values in the 0-1 range, we can anchor the bars at bottom=1 and reduce their height by 1 to keep the top of each bar aligned with your original data values. This way, the 0-1 area stays empty, and your chart's scale remains accurate.

Here's the modified code with key changes highlighted (plus a small cleanup for redundant parameters):

def generate_chart_capability(student, df):
    # canvas sizing
    plt.figure(figsize=(10, 10))
    # plot polar axis (removed redundant projection parameter)
    ax = plt.subplot(111, polar=True)
    ax.set_theta_direction(-1)
    ax.set_theta_offset(np.pi / 1.6)
    ax.set_xticklabels([])
    ax.set_yticklabels([1, 2, 3, 4, 5], size=21)
    # Compute the width of each bar. In total, we have 2*Pi = 360°
    width = 2 * np.pi / len(df.index)
    # Compute the angle each bar is centered on:
    indexes = list(range(1, len(df.index) + 1))
    angles = [element * width for element in indexes]
    # Compute heights: subtract 1 to offset the bottom=1 anchor
    heights = df.Value - 1  # <--- Key change
    # Colour the bars
    c = ['#054998', '#3660B3', '#5579CF', '#7392EB', '#90ADFF', '#ADC8FF']
    # Draw bars anchored at bottom=1 instead of 0
    ax.bar(x=angles, height=heights, width=width, color=c, bottom=1, linewidth=2, edgecolor="white")  # <--- Key change
    ax.set_ylim([0, 5])  # Keep full scale visible (0-1 remains empty)
    # Save the figure
    plt.savefig('app/static/output/capability_chart_' + student + '.png')
    # Close the figure
    plt.clf()

Explanation: #

• bottom=1: Anchors all bars starting at the 1 mark on the polar axis, leaving the 0-1 inner circle blank.
• heights = df.Value - 1: Ensures the top of each bar reaches the original df.Value (since 1 + (df.Value -1) = df.Value), so your data values are accurately represented without shifting.

Solution 2: Overlay a White Circle (For More Control) #

If you want to keep your original bar code unchanged but still block the 0-1 region, you can draw a white circle directly over the center. This is useful if you need to maintain the original bar positioning but hide the inner area:

Add this code right after drawing the bars:

# Draw a white circle to create the donut hole
circle = plt.Circle((0, 0), 1, transform=ax.transData._b, color='white', zorder=10)
ax.add_artist(circle)

Explanation: #

• The circle has a radius of 1 (matching your 0-1 range) and is placed at the center (0,0).
• zorder=10 ensures the circle is drawn on top of the bars, blocking the inner region.
• Adjust the radius value if you need a larger or smaller donut hole.

Either solution will give you the clean donut-style polar bar chart you're looking for!

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