How these notebooks work (start here)¶

Start here · no statistics yet — just how the notebooks work · data: no data yet — just orientation

↑ Course map · Lesson 1: Describing a real dataset →

Brand new to this? You're in the right place. This page has no statistics — it's a five-minute tour of how to drive a notebook so the real lessons feel easy. You do not need to know how to code. You'll mostly read, press one key to run a cell, and look at what appears. Let's go.

1) What is a "notebook"?¶

A notebook is a single scrollable page made of stacked blocks called cells, read top to bottom — like a document where some paragraphs can run. There are just two kinds of cell:

Text cells (like this one) — explanations, headings, and the math. You just read them.
Code cells — the grey boxes. Each contains a little Python. When you run one, the computer does the work and shows the result right underneath.

You got here by double-clicking start-jupyter.bat (a .bat file is just a launcher you double-click). That opened JupyterLab — the free app these lessons run inside — in your web browser, along with a small black window that quietly runs everything behind the scenes. Leave that black window open; closing it stops the course. On the left is a file list of all the lessons; the big area is the lesson you're reading now.

2) How to run a cell — the one move you need¶

Click a grey code cell once (a blue bar appears on its left).
Press Shift + Enter (hold Shift, tap Enter).
The cell runs, its result appears below it, and the cursor hops to the next cell.

On the left of a code cell you'll see [ ]. While it runs it shows [*] (busy), then a number like [1] when it's done — that number just counts the cells you've run, in order.

Try it now on the grey cell just below: click it, then press Shift + Enter.

# This is a code cell. The lines starting with '#' are notes to you (Python ignores them).
import numpy as np
import matplotlib.pyplot as plt

print("✅ You just ran your first cell! See the number that appeared in [ ] on the left.")

# And here's a tiny chart, so you can watch a picture appear:
x = np.arange(1, 11)
fig, ax = plt.subplots(figsize=(6, 3.5))
ax.plot(x, x**2, marker='o', color='#4c72b0')
ax.set_title("A practice chart — you made this appear by running the cell")
ax.set_xlabel("x"); ax.set_ylabel("x × x")
plt.show()

✅ You just ran your first cell! See the number that appeared in [ ] on the left.

No description has been provided for this image

3) Reading the output¶

Whatever a code cell produces shows up directly beneath it:

Printed words or numbers (from print(...)) — like the green check mark above.
Tables — neat grids of data.
Charts — histograms, scatterplots, and the like.

That output is the point. The code is just the recipe; the thing underneath is what it cooked.

4) Run cells in order, top to bottom¶

A lesson builds on itself, so run the cells in order from the top. The very first code cell of every lesson is the setup cell — it loads the tools and the dataset. If a later cell complains that something "is not defined," it almost always means an earlier cell (usually the setup one) hasn't been run yet.

The easy fix for "I think I ran things out of order": in the top menu choose Run → Run All Cells (or Kernel → Restart Kernel and Run All Cells… — the kernel is just the engine that runs your code). That runs everything cleanly from the top.

5) It's safe to experiment¶

You cannot break anything by running or re-running cells. Re-running a cell just recomputes it. The lessons invite you to change a number (a sample size, a random seed, the number of bins) and re-run to see what happens — that's the whole idea of "seeing it happen."

If the notebook ever gets into a weird state, Kernel → Restart Kernel and Run All Cells… is the universal reset button. And nothing you do here touches the original data files (more on that below).

6) Why do the numbers usually stay the same when I re-run?¶

Two reasons, and both are on purpose:

Fixed facts don't change. Many results come from a whole dataset (an average, a count). Re-running just recomputes the same answer.
A fixed "random seed." The simulations use a line like rng = np.random.default_rng(12). That seed makes the "random" draws come out the same every time, so your results match the lesson's. Want to see the randomness wiggle? Change 12 to any other number and re-run — you'll get a different (but similar) result. That's not a bug; it's the lesson.

7) Where does the data come from? (and is it safe?)¶

The real datasets live in the course's raw/ folder. The first time a lesson needs one, it downloads a single copy from a public source and saves it there; after that it loads from your computer, so the course works offline.

Two promises, by design:

Your data is read-only. The course only ever reads those files — it never edits or deletes them.
Nothing of yours is uploaded. Data flows to your machine from public sources; nothing goes out. No account, no sign-in, no cost.

Every dataset's source and licence is written down in raw/SOURCES.md, so any chart can be traced back to where its numbers came from.

8) If something looks broken¶

Red text under a cell = an error. Don't panic — it's usually because an earlier cell wasn't run. Try Kernel → Restart Kernel and Run All Cells….
A cell seems stuck on [*] = it's still working (some simulations take a few seconds). Give it a moment.
The black command window that opened when you launched the course must stay open — it's the engine ("kernel") running your notebooks. Closing it stops the course; re-launch with start-jupyter.bat.
Just want to read, not run? Double-click view-lessons.bat (or open rendered/index.html) for finished pages with every chart already drawn — no setup at all.

You're ready¶

That's the entire skill: read the text, click a grey cell, press Shift + Enter, look below. Everything else is statistics, which the lessons teach from the ground up.

Each lesson explains every idea on three levels — the math (the formula), the meaning (what the number is), and the interpretation (what it says about the world) — and shows it happening on real data.

Next: Lesson 1 — describe your first real dataset (344 Antarctic penguins).

↑ Course map · Next →: Lesson 1: Describing a real dataset