How to Animate the Central Limit Theorem | QuantumSketch
Animate the central limit theorem by drawing sample means for growing n and watching their histogram converge to a bell curve, even from a skewed source.
Animate the central limit theorem (CLT) by repeatedly drawing sample means and watching their histogram converge to a bell curve — even when the source distribution is skewed. That convergence, regardless of the source shape, is the whole surprise.
What the CLT says
The distribution of the sample mean approaches a normal distribution as the sample size n grows, no matter what distribution you sampled from (given finite variance).
A static sentence. The animation makes it click.
The animation, beat by beat
- Show a non-normal source — e.g. a skewed or uniform distribution.
- Draw one sample of size n, compute its mean, drop a dot in a histogram.
- Repeat fast — means pile up into bins.
- Increase n: 5 → 30 → 100. The histogram narrows and symmetrizes.
- Overlay the Gaussian — it fits beautifully, surprising the viewer.
Why growing n matters
| n | Histogram of means | |---|---| | 1 | Looks like the source (skewed) | | 5 | Starting to symmetrize | | 30 | Clearly bell-shaped | | 100 | Tight, near-perfect normal |
Watching the source's skew "wash out" as n grows is the moment students understand the CLT.
Manim building blocks
BarChart / custom rectangles for the histogram, Dot animations for falling samples, axes.plot for the overlaid normal curve, and a counter for n. Driven by a prompt, not by hand-written NumPy.
The prompt
"Sample from a skewed distribution, compute sample means, and build a histogram of those means as n grows 5 → 30 → 100, overlaying the normal curve it approaches."
→ Animate it at quantumsketch.app. See also Animate Bayes' Theorem.
Written by Shihab Shahriar Antor · Shahriar Labs
FAQ
Q.What's the best way to show the central limit theorem visually?
Show many sample means piling up into a histogram. Start with any source distribution — even a lopsided or uniform one — then repeatedly draw a sample of size n, compute its mean, and drop that mean onto a histogram. As you increase n from, say, 1 to 5 to 30 to 100, the histogram of sample means gets narrower and more symmetric, converging to a normal bell curve regardless of the source shape. Animating the means accumulating makes the abstract statement 'the sampling distribution of the mean tends to normal' concrete and surprising.
Q.How do I animate the central limit theorem without statistics software?
Describe it as a prompt: 'Draw samples from a skewed distribution, compute each sample's mean, and build a histogram of those means as the sample size grows from 5 to 30 to 100, showing it approach a normal curve.' QuantumSketch renders this as a narrated Manim animation. Manim can plot the source distribution, animate dots falling into histogram bins, and overlay the fitted Gaussian — all driven by your prompt rather than by you writing NumPy sampling loops and matplotlib code.