📊 Are We Plotting the DataFrame or the Column?

[Rajeev Bagra]

 

Understanding How Pandas and Matplotlib Work Together

When creating visualizations in Python, many learners get confused at this stage:

“We are working with a DataFrame (df), but the boxplot code uses area_m2.
Are we plotting the DataFrame or something else?”

Let’s clarify this properly.

---

✅ The Example We Worked On

import matplotlib.pyplot as plt

fig, ax = plt.subplots()

ax.boxplot(df["area_m2"], vert=False)

ax.set_xlabel("Area [sq meters]")
ax.set_title("Distribution of Home Sizes")

plt.show()

---

🧠 What’s Actually Being Plotted?

We are NOT plotting the entire DataFrame.

We are plotting:

df["area_m2"]

That expression returns a Pandas Series (a single column of data).

Matplotlib only needs array-like numeric values.

So the flow is:

DataFrame → Column (Series) → Matplotlib → Plot

---

📌 Important Concept

Matplotlib does not understand DataFrames directly.

It understands:

• Lists
• NumPy arrays
• Pandas Series (because they behave like arrays)

That’s why this works:

ax.boxplot(df["area_m2"], vert=False)

Because df["area_m2"] behaves like a numeric array.

---

🔎 Why plt.show() Displays the Plot

Another key discussion today:

Even though we use:

ax.boxplot(...)

We still call:

plt.show()

Why?

Because:

• plt (matplotlib.pyplot) manages all active figures.
• plt.show() renders all open figures.
• The boxplot was drawn inside the figure created by plt.subplots().

So plt.show() simply displays whatever figures exist.

---

⚠️ What If There Are Multiple Figures?

If previous figures were created in the same session:

fig1, ax1 = plt.subplots()
fig2, ax2 = plt.subplots()

Then calling:

plt.show()

Will display both.

To prevent this in scripts, use:

plt.close('all')

before creating a new figure.

---

🎨 Two Ways to Plot

1️⃣ Direct Column Extraction (Recommended)

ax.boxplot(df["area_m2"], vert=False)

Clear, readable, and professional.

---

2️⃣ Using an Intermediate Variable

area_m2 = df["area_m2"]
ax.boxplot(area_m2, vert=False)

Both work — but the first is cleaner unless you reuse the variable.

---

🚀 Bonus: Pandas Also Uses Matplotlib

You might see:

df["area_m2"].plot.box()

But here’s the important insight:

👉 Pandas internally calls Matplotlib.

So whether you use Pandas plotting or Matplotlib directly, Matplotlib is still doing the rendering.

---

🎯 Key Takeaways

✔ We are not plotting the entire DataFrame
✔ We are plotting a single column (Series)
✔ Matplotlib needs array-like numeric data
✔ plt.show() displays all active figures
✔ Use plt.close('all') in scripts to avoid multiple plots

---

🧠 Why This Matters

Understanding this distinction helps you:

• Debug plotting errors
• Write cleaner data science code
• Avoid confusion in notebooks
• Think more clearly about data flow

If you truly understand:

DataFrame → Series → Matplotlib → Figure → plt.show()

You’ve moved beyond beginner-level plotting.

 

[Author]

Posts