What is a word cloud?
A word cloud is a picture of a text in which each word's size shows how often it appears. The biggest words are the ones the text keeps coming back to.
Word clouds are also called tag clouds, text clouds or wordles. They are a quick way to get the gist of a long text, a stack of survey answers or a speech without reading every line — and they make a friendly visual for a slide, a classroom wall or a report cover.
How a word cloud is made
Every generator, from 2008's Wordle to CloudShaper, runs the same four steps:
- Split the text into words and count each one, ignoring capital letters.
- Throw away stopwords — “the”, “and”, “of” — which are the most frequent words in almost any text but say nothing about it.
- Turn counts into font sizes. The choice of formula changes the picture more than you'd think.
- Place the words, biggest first, so none overlap.
Step 3: from counts to sizes
Play with the demo: first leave the stopwords in, then remove them and switch between the three scaling formulas.
size = min + (max − min) × n, where n = (count − lowest) ÷ (highest − lowest). Min 14px, max 64px.
| Word | Count | Font size |
|---|---|---|
| the | 42 | 64px |
| and | 31 | 51px |
| students | 18 | 35px |
| of | 17 | 34px |
| reading | 12 | 27px |
| library | 8 | 23px |
| stories | 6 | 20px |
| quiet | 4 | 18px |
| dragons | 2 | 15px |
| maps | 1 | 14px |
“the” and “and” are the biggest words — they tell you nothing about the text. This is why every generator removes stopwords first.
The subtle trap is that readers judge a word by the area of ink, not its height. Double the font size and you roughly quadruple the area, so linear scaling exaggerates the top words. Square-root scaling corrects for that, which is why it's CloudShaper's default (“Balanced” in the Style tab).
Step 4: the spiral
The classic placement method, described by Jonathan Feinberg in his chapter on Wordle in the book Beautiful Visualization (O'Reilly, 2010), is surprisingly simple: put the word at the centre; if it overlaps something, nudge it a little further along a spiral and test again; repeat until it fits.
Words are placed biggest-first, so the most important word always gets the centre.
This demo tests rectangles to keep it easy to follow. Real generators test the actual letter shapes: CloudShaper renders each word to a grid of pixels, so a small word can sit inside the bowl of a big “O” or under the arm of a “T”. That's what makes a cloud dense, and what lets words fill a heart or a letter shape. If you like learning this way — by poking at the mechanism until it clicks — ahaboo has narrated interactive explainers on how everyday things really work.
When a word cloud works — and when it doesn't
Good for
- A first impression of open-ended survey answers
- Sparking discussion in class (“why is unfair so big?”)
- Showing a room what it just said, live
- Vocabulary review and word walls
- Decorative visuals: cards, posters, covers
Weak at
- Precise comparisons — size differences are hard to judge
- Context — “not happy” becomes “happy”
- Fairness between words — long words look heavier
- Synonyms — the same idea split across words (AI mode helps)
- Small texts — a few sentences rarely show a pattern
Visualisation researchers have made these criticisms for years; Marti Hearst's analysis of tag clouds and Jacob Harris's 2011 essay “Word clouds considered harmful” for Nieman Lab are the best-known. The fix isn't to avoid word clouds but to be honest about what they are: an overview. Pair one with the exact numbers from a word frequency counter, or let the AI word cloud generator merge synonyms into themes.
How to make a word cloud
- Open the CloudShaper word cloud generator and paste your text (or a list of “word, number” lines).
- Check the word list: exclude words that dominate for the wrong reason, such as the survey question itself.
- Pick a shape in the Shape tab — or a template — and a palette and font in Style.
- Press “Shuffle layout” until you like the arrangement.
- Download a PNG for slides, or an SVG for print. To put it in a presentation, see word clouds in PowerPoint and Google Slides.
A short history
Sized word lists predate the web: Stanley Milgram's 1976 study of mental maps of Paris, which drew place names at sizes reflecting how often people mentioned them, is often cited as an early example. Tag clouds spread across the web in the mid-2000s on photo-sharing and bookmarking sites, where each clickable tag was sized by how many items carried it. In 2008, Wordle turned the idea into art with tightly packed, mixed-orientation layouts, and that look is what most people now picture when they hear “word cloud”.
Questions people ask
What is a word cloud?
A word cloud (also called a tag cloud or wordle) is a picture of a text in which each word is drawn at a size proportional to how often it appears — or how important it is. The biggest words show at a glance what the text is mostly about.
What is the difference between a word cloud and a tag cloud?
They use the same idea. “Tag cloud” came first, on websites in the 2000s like Flickr and Delicious, where the words were clickable tags sized by how many items used them. “Word cloud” usually means an image made from a body of text.
Why is it sometimes called a wordle?
Wordle was the name of Jonathan Feinberg’s free Java applet, launched in 2008, which popularised the dense, mixed-orientation style of word cloud. It has nothing to do with the 2021 word-guessing game of the same name.
How do you make a word cloud?
Paste your text into a word cloud generator. It counts each word, removes common words like “the”, maps counts to font sizes and places the words on a canvas, biggest first. You then pick a shape, font and colours and download the image.
Are word clouds good for data analysis?
Only as a first impression. People judge word size poorly, long words look more important than short ones, and word clouds lose context such as “not good”. Use a bar chart of the top words when exact comparisons matter.
What does the colour in a word cloud mean?
Usually nothing — colours are decorative. Some tools, including CloudShaper’s AI mode, use colour to show a theme or category; in that case include a legend.