Visualization with Matplotlib & Seaborn
import numpy as np, pandas as pd
import matplotlib.pyplot as plt
plt.rcParams.update({"figure.dpi":110,"axes.grid":True,"grid.alpha":0.25,"font.size":11})
TL, TL2, AM, RD = "#0f766e", "#2dd4bf", "#d97706", "#dc2626"
BASE = "https://raw.githubusercontent.com/johnfisher-ai/Statistics-Data-Science-AI-Visual-Book/main/data/"
fn = "python-for-data-analysis--store-orders.xlsx"
def load(sheet="Orders"):
try: return pd.read_excel("../../data/" + fn, sheet_name=sheet)
except FileNotFoundError: return pd.read_excel(BASE + fn, sheet_name=sheet)
import seaborn as sns
sns.set_theme(style="whitegrid")
orders = load(); products = load("Products")
m = orders.merge(products, on="product_id", how="left")
orders["order_date"] = pd.to_datetime(orders.order_date)
DEMO 1 · The anatomy of a matplotlib figure¶
Every matplotlib chart is a figure (the whole canvas) holding one or more axes (a single plot). You draw on an axis, then label it. Start with the monthly revenue trend as a line.
monthly = orders.groupby(orders.order_date.dt.to_period("M")).revenue.sum()
monthly.index = monthly.index.astype(str)
fig, ax = plt.subplots(figsize=(9, 3.6))
ax.plot(monthly.index, monthly.values, marker="o", color=TL, lw=2)
peak = monthly.idxmax()
ax.scatter([peak], [monthly.max()], color=AM, zorder=5, s=90)
ax.annotate(f"peak {peak}", (peak, monthly.max()), textcoords="offset points", xytext=(0, 10), ha="center", color=AM, fontweight="bold")
ax.set_title("Monthly revenue, 2024"); ax.set_ylabel("dollars"); ax.tick_params(axis="x", rotation=45)
plt.tight_layout(); plt.show()
DEMO 2 · Bar charts for comparing categories¶
Bars compare a value across categories. A horizontal bar chart, sorted, is the clearest way to rank groups like revenue by region.
by_region = orders.groupby("region").revenue.sum().sort_values()
fig, ax = plt.subplots(figsize=(8, 3.4))
ax.barh(by_region.index, by_region.values, color=TL)
for i, v in enumerate(by_region.values):
ax.text(v + 500, i, f"{v:,.0f}", va="center", fontsize=9)
ax.set_title("Revenue by region"); ax.set_xlabel("dollars")
plt.tight_layout(); plt.show()
DEMO 3 · Histograms for the shape of a distribution¶
A histogram bins one numeric column to show its shape. Order revenue is right-skewed: many small orders, a long tail of big ones, which the mean alone would hide.
fig, ax = plt.subplots(figsize=(8, 3.6))
ax.hist(orders.revenue, bins=40, color=TL2, edgecolor="white")
ax.axvline(orders.revenue.mean(), color=RD, ls="--", lw=1.6, label=f"mean {orders.revenue.mean():.0f}")
ax.axvline(orders.revenue.median(), color=AM, ls="--", lw=1.6, label=f"median {orders.revenue.median():.0f}")
ax.set_title("Distribution of order revenue"); ax.set_xlabel("dollars per order"); ax.legend()
plt.tight_layout(); plt.show()
print("Right-skewed: the mean sits above the median, pulled up by the long tail of large orders.")
Right-skewed: the mean sits above the median, pulled up by the long tail of large orders.
DEMO 4 · Scatter plots for relationships¶
A scatter plot puts two numbers on the two axes to reveal how they relate. Color the points by a category to add a third dimension for free.
fig, ax = plt.subplots(figsize=(8, 4))
for cat, sub in m.groupby("category"):
ax.scatter(sub.quantity + np.random.default_rng(0).normal(0, 0.06, len(sub)), sub.revenue,
s=14, alpha=0.5, label=cat)
ax.set_title("Revenue vs quantity, colored by category"); ax.set_xlabel("quantity"); ax.set_ylabel("revenue (dollars)")
ax.legend(fontsize=8, markerscale=1.5)
plt.tight_layout(); plt.show()
DEMO 5 · Seaborn: statistical plots in one line¶
seaborn works directly on a DataFrame and handles the grouping, error bars, and legend for you. A boxplot compares whole distributions across groups; a barplot shows group means with a confidence interval.
fig, ax = plt.subplots(1, 2, figsize=(11, 3.8))
sns.boxplot(data=orders, x="channel", y="revenue", hue="channel", ax=ax[0], palette="crest", legend=False)
ax[0].set_title("Revenue distribution by channel")
sns.barplot(data=m, x="category", y="revenue", hue="category", ax=ax[1], palette="crest", legend=False, errorbar=("ci", 95))
ax[1].set_title("Mean revenue by category (95% CI)"); ax[1].tick_params(axis="x", rotation=20)
plt.tight_layout(); plt.show()
DEMO 6 · A small dashboard: several plots, one figure¶
Real reports combine views. Lay out a two-by-two grid of axes and fill each with a different angle on the same data, a compact dashboard from the pieces you have already built.
fig, ax = plt.subplots(2, 2, figsize=(11, 7))
monthly = orders.groupby(orders.order_date.dt.to_period("M")).revenue.sum(); monthly.index = monthly.index.astype(str)
ax[0,0].plot(monthly.index, monthly.values, marker="o", color=TL); ax[0,0].set_title("Monthly revenue"); ax[0,0].tick_params(axis="x", rotation=45, labelsize=7)
orders.groupby("region").revenue.sum().sort_values().plot.barh(ax=ax[0,1], color=TL); ax[0,1].set_title("Revenue by region")
ax[1,0].hist(orders.revenue, bins=40, color=TL2, edgecolor="white"); ax[1,0].set_title("Order revenue distribution")
sns.boxplot(data=orders, x="channel", y="revenue", hue="channel", ax=ax[1,1], palette="crest", legend=False); ax[1,1].set_title("Revenue by channel")
fig.suptitle("Store performance, 2024", fontsize=14, fontweight="bold")
plt.tight_layout(); plt.show()
Wrap-up¶
matplotlib gave you the grammar (figure, axes, plot, label), and the standard chart types each answer a different question: a line for trend, bars for comparison, a histogram for shape, a scatter for relationship. seaborn added statistical plots in a single call, and subplots composed them into a dashboard. Five notebooks in, you can now take a raw file all the way to a chart, which is the entire arc of Python for data analysis.