Data Visualization Basics

Data visualisation is the art of turning numbers into pictures so people can understand them quickly. The basics come down to two things: picking the right chart for the question you are answering, and keeping the design clean so the message is obvious. A good chart makes a pattern jump out; a bad one hides it.

Choosing the right chart

The chart should match the kind of comparison you want to make:

Start by asking "what is the one thing I want the viewer to notice?" The answer points to the chart type.

Plotting in Python

Matplotlib is the foundational plotting library. Here is a simple bar chart:

import matplotlib.pyplot as plt  # the standard plotting library

products = ["Pen", "Notebook", "Eraser"]
revenue = [950, 1200, 100]

plt.bar(products, revenue)          # draw the bars
plt.title("Revenue by Product")     # always label your chart
plt.xlabel("Product")
plt.ylabel("Revenue")
plt.show()                          # display it

And a line chart for change over time:

import matplotlib.pyplot as plt

months = ["Jan", "Feb", "Mar", "Apr"]
sales = [120, 135, 128, 150]

plt.plot(months, sales, marker="o")   # marker shows each data point
plt.title("Monthly Sales")
plt.ylabel("Sales (thousands)")
plt.show()

Core design principles

• Label everything. Title, axis labels, and units. An unlabelled chart is a puzzle.
• Start bar charts at zero. Truncated axes exaggerate differences and mislead.
• Remove clutter. Drop heavy gridlines, 3D effects, and decorative noise.
• Use colour with purpose. Colour should encode meaning, not decorate. Keep it accessible.
• One message per chart. If you are showing five things, make five charts.

Charts vs statistics

Visualisation and statistics work together. A famous lesson in data is that datasets with identical averages can look completely different when plotted — which is exactly why you should always look at your data, not just compute summaries. This is the spirit of exploratory data analysis.

Common mistakes

• Pie charts with many slices. They are hard to compare. A bar chart almost always communicates better.
• Truncated y-axes. Starting a bar chart above zero distorts the visual comparison.
• Too many colours. A rainbow chart is harder to read than a restrained one.
• No labels. Skipping titles and axis labels forces the viewer to guess.

FAQ

Matplotlib or a BI tool? Both have a place. Matplotlib gives full control in code; tools like Power BI and Tableau make interactive dashboards quickly.

When should I use a pie chart? Rarely — only for two or three parts of a whole. Otherwise prefer a bar chart.

How do I make charts interactive? Libraries like Plotly add hover and zoom, but master static charts first.

Keep learning

Clear visuals are what turn analysis into action. Build the skill alongside pandas for analytics and explore the full Data Science & Analytics hub.

Want to learn visualisation with real projects and feedback? Join the waitlist for the Data Science & Analytics course at Infoplanet, Jalgaon.