Data Visualization — Overview

Before you start: general familiarity with the concept of a chart (bar, line, pie) is assumed — no prior design or statistics background needed.

Why This Exists (The Hook)

A table of 10,000 numbers and a well-chosen line chart can contain the exact same information, but only one of them lets a human spot a trend in two seconds. Data visualization exists because the human visual system is extraordinarily fast at perceiving position, length, and color patterns — and extraordinarily slow at scanning raw numbers for the same patterns. Choosing the right chart type (and avoiding a misleading one) is the difference between a stakeholder actually understanding the data and a stakeholder nodding along without really seeing it.

Analogy — Think of chart selection like choosing the right unit of measurement, not just decoration. You wouldn't measure a room's temperature in kilometers — the unit has to match what you're actually measuring. A pie chart "measures" part-of-whole relationships; a line chart "measures" change over time; a scatter plot "measures" correlation between two things. Using a pie chart to show a trend over 12 months is the visual equivalent of measuring temperature in kilometers — technically a number gets displayed, but it answers the wrong question.

Try it (2 minutes) — Reason through why "start the y-axis at zero for bar charts" is a hard rule, without looking anything up: a bar chart's core visual signal is bar length — a bar twice as tall reads as "twice as much." If the y-axis starts at, say, 90 instead of 0, and one bar is at 92 while another is at 96, that small 4-point difference would visually stretch to look enormous (the shorter bar might appear half the height of the taller one). What false impression does that create about the actual size of the difference, and why would a line chart (where the signal is the slope, not the length from a zero baseline) have a different rule?

What is Data Visualization?

Data visualization is the graphical representation of information and data using charts, graphs, maps, and dashboards. The goal is to make complex data patterns immediately understandable.

Why it matters for Data Analysts:

Core Chart Types and When to Use Each

|-----------|----------|-----------|

The Grammar of Effective Visualizations

1. Choose the right encoding

2. Reduce cognitive load

3. Data-to-ink ratio (Edward Tufte)

Remove everything that doesn't encode information:

Python Libraries for Visualization

# Matplotlib — low-level, full control
import matplotlib.pyplot as plt
fig, ax = plt.subplots(figsize=(10, 6))
ax.bar(categories, values, color='steelblue')
ax.set_title('Monthly Revenue', fontsize=14, fontweight='bold')
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
plt.tight_layout()
plt.savefig('revenue.png', dpi=150)

# Seaborn — statistical visualization
import seaborn as sns
sns.heatmap(df.corr(), annot=True, cmap='coolwarm', fmt='.2f')

# Plotly — interactive charts
import plotly.express as px
fig = px.scatter(df, x='ad_spend', y='revenue', color='channel',
                 size='conversions', hover_data=['campaign'])
fig.show()

# Pandas built-in plotting
df.groupby('month')['revenue'].sum().plot(kind='bar', figsize=(12,5))

Key Principles for Dashboard Design