Welcome

Welcome to Seaborn Foundations: Visualizing Statistical Data.

You already know some Python. You can wrangle a DataFrame with pandas — filter rows, group by a column, compute a mean. You know what an x-axis and a y-axis are. What you want now is to turn columns of numbers into pictures that reveal what the data is actually doing — and to know which picture to reach for, and why.

That is exactly what this course is about.

Why Seaborn?

You could draw every chart in this course with raw matplotlib. But you would spend most of your energy on plumbing — looping over groups, positioning bars, building legends by hand — instead of on seeing.

Seaborn flips that around. You hand it a tidy table and describe your chart in the vocabulary of statistics:

"Put total_bill on x, tip on y, color by time, and split into one panel per day."

Seaborn does the grouping, the aggregation, the color mapping, the legend, and sensible defaults — so the code you write looks like the question you are asking.

This is the mental shift the course is built around: you describe what you want to see; Seaborn figures out how to draw it.

What this course focuses on

This is a statistical visualization course. We care about building intuition: what each chart reveals, what it hides, and when it breaks.

This course is about:

• The principles of statistical data visualization and exploratory data analysis (EDA)
• Tidy data — the table shape Seaborn is built for
• Relational plots (scatter, line), distribution plots (histogram, KDE, ECDF, rug), and categorical plots (bar, count, box, violin, boxen, strip, swarm)
• Regression and trend visualization, correlation heatmaps, and multi-panel grids (facets, pair plots, joint plots)
• Aesthetics that aid communication — themes, contexts, and color palettes — and the judgment to tell a clear visual story

This course is not about:

• Low-level matplotlib scripting (we use it only for light touch-ups)
• Machine-learning diagnostic plots (ROC curves, partial dependence, ...)
• Interactive web charting (Plotly, Bokeh, Dash) or BI dashboards
• Geographic / GIS maps, 3-D plots, or real-time streaming charts
• Graphic-design theory beyond what a working analyst needs

We borrow ideas from those worlds only when they sharpen our intuition about Seaborn.

How the interactive pages work

Every code block on these pages runs right here in your browser — there is nothing to install, and Seaborn, pandas, and the example datasets are already available. Edit any snippet, run it, break it, and fix it. That is how the intuition sticks.

You will meet three kinds of widgets:

1. Executable code blocks. Change a parameter — hue, bins, kind — and watch the chart respond.
2. Challenge cards. Small, checked tasks where you write the code. Hidden tests tell you the moment you have it right.
3. Multiple-choice questions. Quick concept checks. The more foundational a page is, the more of these you will find — getting the mental model right early saves hours later.

A taste of where we are going

Here is a complete, multi-panel Seaborn figure. Do not worry about the details yet — by the end of this course every line will feel obvious, and you will be able to say what each one is for.

Notice what you did not have to do: no loop over species, no manual color assignment, no legend code, no two-subplot scaffolding. You named the columns and the roles they should play — x, y, hue, col — and Seaborn did the rest. That is the whole game.

Ready? Let's start with the most important question of all: why bother turning data into pictures at all.