04-visualization

Visualization with Matplotlib

Objective

Learn how to turn data into charts and graphs using matplotlib, Python's most popular plotting library. You'll create line plots, bar charts, scatter plots, histograms, pie charts, and more.

Concepts Covered

• What matplotlib is and why visualization matters
• The pyplot interface
• Line plots, bar charts, scatter plots, histograms, and pie charts
• Customizing plots with titles, labels, legends, colors, and styles
• Creating multiple subplots in one figure
• Saving figures to files
• Built-in themes with plt.style.use()
• Brief overview of alternative libraries (seaborn, plotly)

Prerequisites

• Basic Python (lists, loops, functions)
• Familiarity with data concepts (the earlier lessons in this section help, but aren't strictly required)
• matplotlib installed (pip install matplotlib)

Lesson

Why Visualization Matters

Numbers in a list are hard to interpret. A chart tells the story instantly. Visualization helps you:

• Spot patterns — trends, clusters, outliers jump out visually
• Communicate findings — a chart is worth a thousand rows of data
• Debug your data — weird data shows up fast when you plot it
• Explore before analyzing — always plot your data before running fancy statistics

What is Matplotlib?

Matplotlib is Python's foundational plotting library. It's been around since 2003 and virtually every other Python visualization tool builds on top of it. It can produce publication-quality charts in a wide variety of formats.

To install it:

pip install matplotlib

The pyplot Interface

Matplotlib has a lot of layers, but you'll almost always use the pyplot interface. By convention, everyone imports it as plt:

import matplotlib.pyplot as plt

The basic workflow is always the same:

1. Create some data
2. Call a plotting function (plt.plot(), plt.bar(), etc.)
3. Customize it (titles, labels, colors)
4. Show it (plt.show()) or save it (plt.savefig())

Line Plots: plt.plot()

The most basic chart. Great for showing trends over time:

days = [1, 2, 3, 4, 5, 6, 7]
temperatures = [72, 75, 71, 68, 74, 78, 82]

plt.plot(days, temperatures)
plt.title("Weekly Temperature")
plt.xlabel("Day")
plt.ylabel("Temperature (F)")
plt.savefig("line_plot.png")
plt.close()

You can plot multiple lines on the same chart:

plt.plot(days, city_a_temps, label="City A")
plt.plot(days, city_b_temps, label="City B")
plt.legend()  # Shows which line is which

Bar Charts: plt.bar() and plt.barh()

Perfect for comparing categories:

languages = ["Python", "JavaScript", "Java", "C++"]
popularity = [30, 25, 20, 15]

plt.bar(languages, popularity)
plt.title("Language Popularity")
plt.ylabel("Popularity (%)")
plt.savefig("bar_chart.png")
plt.close()

Use plt.barh() for horizontal bars — handy when your category labels are long:

plt.barh(languages, popularity)

Scatter Plots: plt.scatter()

Shows the relationship between two variables. Each point is one observation:

heights = [62, 65, 68, 70, 72, 74, 67, 71]
weights = [130, 150, 165, 180, 190, 200, 155, 175]

plt.scatter(heights, weights)
plt.title("Height vs Weight")
plt.xlabel("Height (inches)")
plt.ylabel("Weight (lbs)")
plt.savefig("scatter_plot.png")
plt.close()

You can control the size and color of each point with the s and c parameters:

plt.scatter(heights, weights, s=100, c="coral", alpha=0.7)

Histograms: plt.hist()

Shows the distribution of a single variable — how often values fall into different ranges (bins):

scores = [88, 92, 78, 95, 67, 82, 90, 85, 73, 91, 88, 76, 94, 81, 87]

plt.hist(scores, bins=5, edgecolor="black")
plt.title("Test Score Distribution")
plt.xlabel("Score")
plt.ylabel("Number of Students")
plt.savefig("histogram.png")
plt.close()

The bins parameter controls how many bars you get. More bins = more detail.

Pie Charts: plt.pie()

Shows parts of a whole. Use sparingly — bar charts are usually easier to read, but pie charts work well for simple breakdowns:

activities = ["Sleep", "Work", "Commute", "Leisure", "Eating"]
hours = [8, 9, 1.5, 4, 1.5]

plt.pie(hours, labels=activities, autopct="%1.0f%%", startangle=90)
plt.title("How I Spend My Day")
plt.savefig("pie_chart.png")
plt.close()

• autopct adds percentage labels to each slice
• startangle rotates the starting position

Customizing Your Plots

Matplotlib gives you control over almost everything. Here are the most common tweaks:

# Colors — use names, hex codes, or RGB tuples
plt.plot(x, y, color="coral")
plt.plot(x, y, color="#FF6347")
plt.plot(x, y, color=(0.2, 0.4, 0.6))

# Line styles
plt.plot(x, y, linestyle="--")    # Dashed
plt.plot(x, y, linestyle=":")     # Dotted
plt.plot(x, y, linestyle="-.")    # Dash-dot

# Markers — highlight data points
plt.plot(x, y, marker="o")       # Circles
plt.plot(x, y, marker="s")       # Squares
plt.plot(x, y, marker="^")       # Triangles

# Shorthand — combine color, marker, and line in one string
plt.plot(x, y, "ro--")  # Red circles with dashed line

# Titles and labels
plt.title("My Chart", fontsize=16, fontweight="bold")
plt.xlabel("X Axis", fontsize=12)
plt.ylabel("Y Axis", fontsize=12)

# Grid
plt.grid(True, linestyle="--", alpha=0.7)

# Annotations — point out something specific
plt.annotate("Peak!", xy=(5, 95), fontsize=12,
             arrowprops=dict(arrowstyle="->"))

Subplots: Multiple Plots in One Figure

When you want to compare several charts side by side, use plt.subplots():

fig, axes = plt.subplots(2, 2, figsize=(10, 8))

# axes is a 2x2 array — access each subplot by row and column
axes[0, 0].plot(x, y)
axes[0, 0].set_title("Line Plot")

axes[0, 1].bar(categories, values)
axes[0, 1].set_title("Bar Chart")

axes[1, 0].scatter(x, y)
axes[1, 0].set_title("Scatter Plot")

axes[1, 1].hist(data, bins=10)
axes[1, 1].set_title("Histogram")

plt.tight_layout()  # Prevents overlap
plt.savefig("subplots.png")
plt.close()

Notice that with subplots, you call methods on the axes objects (like axes[0, 0].set_title()) instead of on plt directly. This trips people up at first — just remember: plt controls the whole figure, axes control individual subplots.

Saving Figures: plt.savefig()

You can save to PNG, PDF, SVG, and more. The format is determined by the file extension:

plt.savefig("chart.png")              # PNG (raster, good for web)
plt.savefig("chart.pdf")              # PDF (vector, good for papers)
plt.savefig("chart.svg")              # SVG (vector, good for web)
plt.savefig("chart.png", dpi=300)     # Higher resolution
plt.savefig("chart.png", bbox_inches="tight")  # Trim whitespace

Always call plt.savefig() before plt.show() or plt.close() — those functions clear the figure.

Built-in Themes: plt.style.use()

Matplotlib comes with several built-in themes that change the overall look of your plots:

plt.style.use("ggplot")           # R-inspired style
plt.style.use("seaborn-v0_8")    # Clean and modern
plt.style.use("dark_background")  # Dark mode
plt.style.use("fivethirtyeight")  # News-style charts

You can list all available styles with:

print(plt.style.available)

To temporarily use a style without changing your default:

with plt.style.context("ggplot"):
    plt.plot(x, y)
    plt.savefig("styled.png")

Alternative Libraries Worth Knowing

Matplotlib is the foundation, but there are libraries that build on it or take a different approach:

• seaborn — Built on top of matplotlib. Makes statistical charts (heatmaps, violin plots, pair plots) much easier. Great default styles. Install with pip install seaborn.

• plotly — Creates interactive charts that you can zoom, pan, and hover over. Runs in the browser. Great for dashboards and presentations. Install with pip install plotly.

• pandas built-in plotting — If you have a DataFrame, you can call df.plot() directly. It uses matplotlib under the hood but saves you some typing.

You don't need to learn these right now — matplotlib is the essential one. But when you find yourself writing a lot of boilerplate for a specific chart type, check if seaborn or plotly makes it easier.

Code Example

Check out example.py for a complete working example that demonstrates everything above.

Exercises

Try the practice problems in exercises.py to test your understanding.

Key Takeaways

• matplotlib is Python's most popular plotting library — import matplotlib.pyplot as plt
• The workflow is always: create data, plot it, customize it, save/show it
• plt.plot(), plt.bar(), plt.scatter(), plt.hist(), and plt.pie() cover most chart types
• Customize with plt.title(), plt.xlabel(), plt.ylabel(), plt.legend(), and plt.grid()
• Use plt.subplots() to put multiple charts in one figure
• Use plt.savefig() to export your charts to PNG, PDF, SVG, etc.
• plt.style.use() gives you quick theme changes with zero effort
• Always call plt.close() when you're done with a figure to free memory
• seaborn and plotly are great next steps once you're comfortable with matplotlib