How to create a box plot in Matplotlib

Box plots make several numeric distributions comparable in one compact figure. In Matplotlib, a box plot shows each group's median, quartile range, whiskers, and outlier points, which helps when values need more context than a single average.

Pass one numeric sequence per category to Axes.boxplot(). The sequences stay in their supplied order, so each tick_labels entry must match the data sequence at the same position.

Set patch_artist=True when the boxes need fill colors, because that option returns each box as a patch that can be styled after plotting. The finished figure should retain three category labels and expose unusually distant values as individual outlier markers.

1. Create box_plot_create.py containing the imports, grouped latency samples, matching labels, and box palette.

   box_plot_create.py

   from pathlib import Path
    
   import matplotlib.pyplot as plt
   import numpy as np
    
   rng = np.random.default_rng(42)
   latency_ms = [
       rng.normal(95, 12, 80),
       rng.normal(110, 18, 80),
       rng.normal(132, 25, 80),
   ]
   labels = ["API", "Batch", "Search"]
   colors = ["#d8ecff", "#e6f4d7", "#ffe2c2"]

   Each array in latency_ms corresponds to the labels string at the same position.

1. Append the figure and box-plot construction below the data definitions.

   box_plot_create.py

   fig, ax = plt.subplots(figsize=(7, 4), layout="constrained")
   result = ax.boxplot(
       latency_ms,
       tick_labels=labels,
       patch_artist=True,
       medianprops={"color": "black", "linewidth": 1.4},
       boxprops={"edgecolor": "#3973ac"},
   )

   The tick_labels argument supplies one label per data sequence. Setting patch_artist=True makes the returned boxes artists fillable.

2. Add the box colors, axis text, title, and horizontal reference grid after the boxplot() call.

   box_plot_create.py

   for box, color in zip(result["boxes"], colors):
       box.set_facecolor(color)
    
   ax.set_ylabel("Latency (ms)")
   ax.set_title("Service latency by workload")
   ax.grid(axis="y", linestyle=":", alpha=0.5)

3. Finish the script with PNG export and plot-object checks.

   box_plot_create.py

   output = Path("box-plot-create.png")
   fig.savefig(output, dpi=150)
    
   tick_text = ", ".join(label.get_text() for label in ax.get_xticklabels())
   print(f"boxes: {len(result['boxes'])}")
   print(f"tick labels: {tick_text}")
   print(f"saved: {output}")

4. Run box_plot_create.py from the directory that should receive the exported image.

   $ python3 box_plot_create.py
   boxes: 3
   tick labels: API, Batch, Search
   saved: box-plot-create.png

5. Inspect the exported PNG file type and pixel dimensions.

   $ file box-plot-create.png
   box-plot-create.png: PNG image data, 1050 x 600, 8-bit/color RGBA, non-interlaced

6. Confirm box-plot-create.png shows three labeled distributions with median lines, whiskers, and outlier markers.