Python实现堆叠点直方图（以堆叠点列替代柱状条）

Got it! I’ve helped folks build this exact visualization before—it’s a fantastic way to show distribution while keeping visibility on individual data points, instead of just aggregated bars. Let’s walk through a complete implementation using matplotlib and numpy (the go-to tools for custom Python visualizations):

Step 1: Core Concept #

Instead of drawing a bar for each histogram bin (where height = count), we’ll:

• Split our data into bins just like a regular histogram
• For each bin, plot individual points stacked vertically—each point represents one data point in that bin
• Align points neatly so they don’t overlap (unless you want intentional jitter for dense bins)

Step 2: Full Code Implementation #

import numpy as np
import matplotlib.pyplot as plt

# ----------------------
# 1. Prepare your data (replace this with your own dataset!)
# ----------------------
np.random.seed(42)  # For reproducibility
data = np.concatenate([
    np.random.normal(loc=2, scale=0.8, size=50),
    np.random.normal(loc=5, scale=1.2, size=70),
    np.random.normal(loc=8, scale=0.6, size=40)
])

# ----------------------
# 2. Set up histogram bins
# ----------------------
bin_edges = np.arange(0, 10, 0.5)  # Customize bin width as needed
bin_counts, _ = np.histogram(data, bins=bin_edges)
bin_centers = (bin_edges[:-1] + bin_edges[1:]) / 2  # X-position for each bin's points

# ----------------------
# 3. Calculate positions for each stacked point
# ----------------------
x_coords = []
y_coords = []
for bin_idx, count in enumerate(bin_counts):
    if count == 0:
        continue
    # X: repeat the bin center for every point in the bin
    x_coords.extend([bin_centers[bin_idx]] * count)
    # Y: stack points vertically (start at 0, increment by 1 per point)
    y_coords.extend(range(count))

# ----------------------
# 4. Plot the stacked dots
# ----------------------
plt.figure(figsize=(10, 6))
plt.scatter(x_coords, y_coords, s=50, color="#2c3e50", alpha=0.8, edgecolor="white")

# ----------------------
# 5. Customize the plot to match your desired style
# ----------------------
plt.xlabel("Value", fontsize=12)
plt.ylabel("Count (Stacked Points)", fontsize=12)
plt.title("Stacked Dot Histogram", fontsize=14, pad=20)
plt.xlim(bin_edges[0] - 0.2, bin_edges[-1] + 0.2)
plt.ylim(-0.5, max(bin_counts) + 0.5)

# Optional: Add grid lines for readability
plt.grid(axis='y', linestyle='--', alpha=0.7)

# Remove top/right spines for cleaner look
plt.gca().spines[['top', 'right']].set_visible(False)

plt.tight_layout()
plt.show()

Step 3: Key Customizations #

• Bin Width: Adjust bin_edges to make bins wider/narrower (e.g., np.arange(0,10,1) for wider bins)
• Point Style: Change s (size), color, alpha, or marker (e.g., marker='o' or marker='s')
• Jitter for Dense Bins: If you have bins with hundreds of points, add small random noise to X positions to avoid overcrowding:

  x_coords = [x + np.random.uniform(-0.1, 0.1) for x in x_coords]
  

• Color Coding: If you want to group data by category, assign different colors to subsets of points (just split your data into groups and plot each group separately)

Step 4: Matching Your Example Visualization #

Since you mentioned having an example image, you can tweak the above code to match:

• Adjust bin positions to align with your example
• Change point colors/sizes to match the style
• Add any annotations or labels present in your example

内容的提问来源于stack exchange，提问作者jasikevicius23