Visualization Basics

8 examples to get you started with Visualization - 6 basic and 2 intermediate.

Prerequisites

• Python 3.14.0 with matplotlib, seaborn, and pandas 2.2+.
• Quick setup: uv pip install matplotlib seaborn pandas plotly

Basic Examples

1. Line Chart for Trends

Show change over ordered time or categories.

import matplotlib.pyplot as plt
 
months = ["Jan", "Feb", "Mar", "Apr"]
revenue = [120, 135, 128, 150]
plt.plot(months, revenue, marker="o")
plt.ylabel("Revenue ($k)")
plt.title("Monthly Revenue")
plt.tight_layout()
plt.savefig("trend.png", dpi=150)
plt.close()

• Line charts need ordered x-axis - not arbitrary categories.
• Markers help sparse series; dense series can omit them.
• Always call plt.close() in scripts to free memory.

Related: matplotlib - figures and axes

2. Bar Chart for Comparisons

Compare discrete categories side by side.

import matplotlib.pyplot as plt
 
regions = ["East", "West", "Central"]
totals = [420, 380, 290]
plt.bar(regions, totals, color=["#2c7bb6", "#abd9e9", "#fdae61"])
plt.ylabel("Total revenue")
plt.tight_layout()

• Bar length encodes magnitude - start y-axis at zero for honesty.
• Sort bars by value when order is not intrinsic.
• Limit categories (~15) or switch to horizontal bars.

Related: seaborn - statistical bar plots

3. Histogram for Distributions

Reveal spread and skew of a numeric variable.

import matplotlib.pyplot as plt
import numpy as np
 
rng = np.random.default_rng(42)
values = rng.normal(100, 15, size=500)
plt.hist(values, bins=20, edgecolor="white")
plt.xlabel("Score")
plt.ylabel("Count")

• Bin count affects story - try 15-30 bins for hundreds of points.
• Density=True compares distributions with different sample sizes.
• Pair with box/violin plots when comparing groups.

Related: seaborn - histplot and KDE

4. Scatter for Relationships

Explore correlation between two numeric variables.

import matplotlib.pyplot as plt
 
ad_spend = [10, 20, 15, 30, 25]
revenue = [120, 180, 150, 240, 210]
plt.scatter(ad_spend, revenue, alpha=0.7)
plt.xlabel("Ad spend ($k)")
plt.ylabel("Revenue ($k)")

• Alpha transparency helps overplotting clusters.
• Add regression line via seaborn regplot when trend is the message.
• Label outliers when they drive decisions.

Related: seaborn - scatterplot with hues

5. seaborn Statistical Defaults

Get readable themes and color palettes quickly.

import seaborn as sns
import matplotlib.pyplot as plt
import pandas as pd
 
df = pd.DataFrame({"region": ["East", "West"] * 5, "revenue": range(10)})
sns.set_theme(style="whitegrid")
sns.barplot(data=df, x="region", y="revenue", estimator=sum, errorbar=None)
plt.tight_layout()

• data= API avoids manual axis juggling.
• estimator and errorbar control aggregation semantics.
• Reset theme after if mixing raw matplotlib styles.

Related: seaborn - EDA workflows

6. Save Publication-Quality Output

Control dpi, bbox, and vector formats.

import matplotlib.pyplot as plt
 
fig, ax = plt.subplots(figsize=(6, 4))
ax.plot([1, 2, 3], [1, 4, 9])
fig.savefig("chart.pdf", bbox_inches="tight")  # vector for print
fig.savefig("chart.png", dpi=300, bbox_inches="tight")
plt.close(fig)

• PDF/SVG scale cleanly in reports; PNG needs dpi (150 screen, 300 print).
• bbox_inches="tight" trims excess margins.
• Explicit fig, ax pattern scales to subplots.

Related: matplotlib - figure lifecycle

Intermediate Examples

7. Interactive Plotly Hover

Explore dense points with tooltips in the browser.

import plotly.express as px
import pandas as pd
 
df = pd.DataFrame({"x": range(20), "y": [i**1.5 for i in range(20)], "cat": ["A", "B"] * 10})
fig = px.scatter(df, x="x", y="y", color="cat", hover_data=["x", "y"])
fig.write_html("scatter.html")

• HTML exports embed in dashboards without a server.
• Plotly shines for cross-filter linked views.
• For static reports, prefer matplotlib/seaborn.

Related: Plotly - interactive charts

8. Declarative Altair Chart

Compose grammar-of-graphics specs that stay readable.

import altair as alt
import pandas as pd
 
df = pd.DataFrame({"month": ["Jan", "Feb", "Mar"], "revenue": [120, 135, 128]})
chart = alt.Chart(df).mark_line(point=True).encode(x="month", y="revenue")
chart.save("revenue.json")  # Vega-Lite spec

• Altair enforces tidy data - one row per observation.
• JSON specs version-control well in git.
• Large data needs aggregation before Altair.

Related: Altair - declarative encoding

Stack versions: This page was written for Python 3.14.0 (stable 3.14, maintenance 3.13), FastAPI 0.115+, Django 5.2, Flask 3.1, Pydantic 2, PyTorch 2.6+, pandas 2.2+, Polars 1.x, ruff 0.9+, and uv 0.6+.