Subplots & styling

One chart answers one question. A review deck usually needs three or four related charts side by side, in a consistent style, exported at the right size. That is what subplots and styling are for.

Many panels in one figure

plt.subplots(nrows, ncols) returns the Figure and an array of Axes.

import matplotlib.pyplot as plt

fig, axes = plt.subplots(1, 2, figsize=(10, 4))

axes[0].hist(df["age"], bins=20, color="#4C78A8", edgecolor="white")
axes[0].set_title("Age distribution")
axes[0].set_xlabel("Age (years)")

counts = df["city"].value_counts().head(6)
axes[1].bar(counts.index, counts.values, color="#F58518")
axes[1].set_title("Customers by city")
axes[1].tick_params(axis="x", rotation=45)

fig.suptitle("Customer profile — Q2 FY26", fontsize=14)
fig.tight_layout()
fig.savefig("profile.png", dpi=200, bbox_inches="tight")

With more than one row, axes is 2-D: use axes[0, 1], or loop over axes.flat when you build one panel per category.

fig, axes = plt.subplots(2, 3, figsize=(12, 6), sharey=True)
for ax, (city, part) in zip(axes.flat, df.groupby("city")):
    ax.plot(part["month"], part["sales"])
    ax.set_title(city)
for ax in axes.flat[len(df["city"].unique()):]:
    ax.set_visible(False)      # hide leftover empty panels

sharey=True forces one common y scale, which is what makes small multiples comparable. Without it, a city selling ₹2 lakh and one selling ₹40 lakh look identical.

Styling once, not chart by chart

plt.style.use("seaborn-v0_8-whitegrid")
plt.rcParams.update({
    "figure.dpi": 110,
    "font.size": 11,
    "axes.titlesize": 13,
    "axes.spines.top": False,
    "axes.spines.right": False,
})

plt.style.use() picks a whole look; rcParams overrides individual defaults for every chart in the session. Keep this block in one module so every analyst in the team produces charts that match. Per-chart tweaks still work:

ax.spines[["top", "right"]].set_visible(False)
ax.legend(title="Store", frameon=False, loc="upper left")
ax.set_ylim(0, None)          # bars and areas must start at zero

Saving for different destinations

In a loop, close figures with plt.close(fig) so memory does not creep up when you export 200 store charts.

Key takeaways

• fig, axes = plt.subplots(r, c) first; then treat each Axes as its own little chart.
• sharex / sharey make panels comparable; fig.suptitle gives the group one headline.
• Set style globally with rcParams, not by repeating colour arguments everywhere.
• Always bbox_inches="tight" when saving, and save before plt.show().