TECH **Matplotlib Basics: Figure, Axes, Subplots, Styles**

Matplotlib Basics: Figure, Axes, Subplots, Styles

A Hyper-Practical, Zero-Fluff Study Guide for Data Scientists

1. What This Is & Why It Matters

Matplotlib is the foundation of Python data visualization. If you’ve ever generated a plot in pandas (df.plot()), you’ve used Matplotlib under the hood. But when you need custom layouts, multiple subplots, or fine-grained styling, you must work directly with Matplotlib’s core objects: Figure and Axes.

Why This Matters in Production

• Broken dashboards: If you don’t understand Figure vs. Axes, your plots will overlap, misalign, or fail to render in Jupyter/Streamlit.
• Slow reports: Poorly structured subplots waste CPU cycles and memory, especially when generating hundreds of plots (e.g., in a batch ETL job).
• Unprofessional outputs: Default styles look dated. Custom styles (fonts, colors, grids) make your work stand out in client presentations or internal reviews.

Real-World Scenario

You’re building a real-time monitoring dashboard for a logistics company. The dashboard must show: 1. A line plot of daily package volumes (2020–2024).
2. A bar plot of delivery delays by region.
3. A pie chart of shipping method distribution.

Problem: If you use plt.plot() for each chart, they’ll stack vertically in one figure, wasting space. If you use subplots incorrectly, the pie chart will stretch into an oval. Solution: Master Figure, Axes, and subplots to control layout precisely.

2. Core Concepts & Components

3. Step-by-Step Hands-On: Build a Multi-Panel Dashboard

Prerequisites

• Python 3.8+ with matplotlib and pandas installed: bash pip install matplotlib pandas
• A dataset (we’ll use a synthetic one for this guide).

Task

Create a 3-panel dashboard with: 1. A line plot of daily sales (2020–2024).
2. A bar plot of sales by region.
3. A pie chart of product categories.

Step 1: Generate Synthetic Data

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

# Generate synthetic data
np.random.seed(42)
dates = pd.date_range("2020-01-01", "2024-12-31", freq="D")
sales = np.cumsum(np.random.normal(100, 10, len(dates))) + 1000
regions = ["North", "South", "East", "West"]
region_sales = {r: np.random.randint(500, 2000) for r in regions}
categories = ["Electronics", "Clothing", "Furniture", "Groceries"]
category_sales = [35, 25, 20, 20]

# Create DataFrames
df_sales = pd.DataFrame({"Date": dates, "Sales": sales})
df_regions = pd.DataFrame({"Region": regions, "Sales": region_sales.values()})
df_categories = pd.DataFrame({"Category": categories, "Sales": category_sales})

Step 2: Create a Figure with Subplots

# Create a 2x2 grid of subplots (we'll use 3 of them)
fig, axes = plt.subplots(nrows=2, ncols=2, figsize=(12, 8))

# Flatten axes for easy iteration (axes is a 2x2 array)
axes = axes.flatten()

# Line plot (top-left)
axes[0].plot(df_sales["Date"], df_sales["Sales"], color="tab:blue")
axes[0].set_title("Daily Sales (2020–2024)")
axes[0].set_xlabel("Date")
axes[0].set_ylabel("Sales ($)")

# Bar plot (top-right)
axes[1].bar(df_regions["Region"], df_regions["Sales"], color="tab:orange")
axes[1].set_title("Sales by Region")
axes[1].set_ylabel("Sales ($)")

# Pie chart (bottom-left)
axes[2].pie(
    df_categories["Sales"],
    labels=df_categories["Category"],
    autopct="%1.1f%%",
    startangle=90,
    colors=["tab:green", "tab:red", "tab:purple", "tab:brown"]
)
axes[2].set_title("Sales by Category")

# Hide the unused subplot (bottom-right)
axes[3].axis("off")

# Adjust layout to prevent overlap
plt.tight_layout()

# Save the figure
plt.savefig("sales_dashboard.png", dpi=300, bbox_inches="tight")
plt.show()

Expected Output

(Example output; actual plot will vary slightly.)

Step 3: Verify Success

1. Check the saved file sales_dashboard.png:
   bash
ls -lh sales_dashboard.png # Should be ~200–500 KB
2. Open the file to confirm:
3. No overlapping labels.
4. Consistent font sizes.
5. All 3 plots are visible.

4. ? Production-Ready Best Practices

Layout & Performance

• Use plt.subplots() instead of plt.subplot(): Cleaner code and better control over Axes.
• Close figures explicitly: In long-running scripts, call plt.close(fig) to free memory.
• Set figsize early: Adjust figsize=(width, height) to match your output (e.g., (10, 6) for reports).
• Use tight_layout(): Prevents label overlap. For fine control, use plt.subplots_adjust().

Styling

• Enforce team-wide styles: Override rcParams at the start of scripts: python plt.rcParams.update({
"font.size": 12,
"axes.titlesize": 14,
"axes.labelsize": 12,
"xtick.labelsize": 10,
"ytick.labelsize": 10,
"figure.titlesize": 16,
"figure.figsize": (10, 6) })
• Use plt.style.use(): Apply consistent themes (e.g., seaborn, ggplot).
• Avoid default colors: Use tab10 or custom palettes for accessibility.

Saving Figures

• Always set dpi=300 for reports: Low DPI (default=100) looks pixelated.
• Use bbox_inches="tight": Prevents cropped labels.
• Prefer vector formats (PDF/SVG) for reports: Avoids pixelation when zooming.

Automation

• Loop over subplots: For dynamic dashboards, use for ax in axes: to iterate.
• Reuse figures: Clear axes with ax.clear() instead of creating new figures.

5. ⚠️ Common Mistakes & Traps

6. ? Exam/Certification Focus

Typical Question Patterns

1. Subplot Layouts:
2. "How do you create a 2x3 grid of subplots?"
   Answer: fig, axes = plt.subplots(2, 3, figsize=(12, 8)).

3. "How do you hide an unused subplot?"
   Answer: axes[5].axis("off").

4. Styling:

5. "How do you apply the ggplot style to all plots?"
   Answer: plt.style.use("ggplot").

6. "How do you set the default font size for all plots?"
   Answer: plt.rcParams["font.size"] = 12.

7. Saving Figures:

8. "How do you save a figure as a 300 DPI PNG?"
   Answer: plt.savefig("plot.png", dpi=300, bbox_inches="tight").

Key ⚠️ Trap Distinctions

• plt.plot() vs. ax.plot():
• plt.plot(): Works on the "current" figure/axes (unpredictable in subplots).
• ax.plot(): Explicitly plots on a specific Axes object (recommended).
• plt.subplots() vs. plt.subplot():
• plt.subplots(): Creates all subplots at once (cleaner for grids).
• plt.subplot(): Adds one subplot at a time (messy for complex layouts).

Scenario-Based Question

"You need to create a dashboard with 4 plots (2 rows, 2 columns). The bottom-right plot is unused. How do you hide it?" Answer:

fig, axes = plt.subplots(2, 2, figsize=(10, 8))
axes[1, 1].axis("off")  # Hide the bottom-right subplot

7. ? Hands-On Challenge

Challenge

Create a 2x2 subplot grid where: 1. Top-left: Line plot of y = sin(x) (0 to 2π).
2. Top-right: Scatter plot of y = cos(x) vs. x.
3. Bottom-left: Histogram of 1000 random normal values.
4. Bottom-right: Hide the subplot.

Constraints: - Use plt.subplots().
- Apply the seaborn style.
- Save the figure as trig_plots.png (300 DPI).

Solution

import numpy as np
import matplotlib.pyplot as plt

# Apply style
plt.style.use("seaborn")

# Create subplots
fig, axes = plt.subplots(2, 2, figsize=(10, 8))
x = np.linspace(0, 2 * np.pi, 100)

# Top-left: sin(x)
axes[0, 0].plot(x, np.sin(x), color="tab:blue")
axes[0, 0].set_title("sin(x)")

# Top-right: cos(x) scatter
axes[0, 1].scatter(x, np.cos(x), color="tab:orange", s=10)
axes[0, 1].set_title("cos(x)")

# Bottom-left: histogram
axes[1, 0].hist(np.random.normal(0, 1, 1000), bins=30, color="tab:green")
axes[1, 0].set_title("Normal Distribution")

# Hide bottom-right
axes[1, 1].axis("off")

# Adjust layout and save
plt.tight_layout()
plt.savefig("trig_plots.png", dpi=300, bbox_inches="tight")
plt.show()

Why It Works: - plt.subplots(2, 2) creates a 2x2 grid.
- axes[i, j] accesses subplots by row/column.
- axis("off") hides the unused subplot.
- tight_layout() prevents label overlap.

8. ? Rapid-Reference Crib Sheet

9. ? Where to Go Next

1. Matplotlib Official Tutorials – Start with "Pyplot" and "Subplots".
2. Seaborn Gallery – See how Matplotlib is used under the hood in Seaborn.
3. Python Data Science Handbook (Jake VanderPlas) – Chapter 4 covers Matplotlib in depth.
4. Real Python – Matplotlib Guide – Practical examples and best practices.