What is a word cloud, and what is it good for?

A word cloud is a type of data visualisation, created from text, where words, phrases or themes are arranged in different sizes and positions within a "cloud" shape. The size, and sometimes the position, of each word represents how often it appears in the text. The most frequent word or phrase is usually the largest, and usually sits in the middle of the cloud.

Word clouds are a kind of tag cloud. A tag cloud is a visual representation of information that uses the size of objects to express their weight or importance; a word cloud is a tag cloud that displays information from text. To make one from your own text, paste it into the free word cloud generator. The preview needs no account.

Where does the text for a word cloud come from?

Anything you can copy and paste text from: a website, a text document, a spreadsheet. Common sources include articles, reviews, press releases, interviews, discussions and the open-ended comments collected in surveys.

• Books and long documents. Our co-founder, Ray Poynter, made a word cloud of A Christmas Carol using just the text file version of the book.
• Market research data, such as what people think of a brand's customer service or how they describe a product.
• Academic papers, research conference talks and presentations, where you want insight into what has been said.
• Video. With good transcription services you can turn a video into words and put the transcript through a word cloud generator to get its key themes.

Put simply, a data file is anything and everything where there is text to be analysed.

Five ways word clouds are useful

1. They save time. A word cloud shows the key terms in a data file without anyone reading the whole file. Reading 5,000 product reviews can take many hours, compared with a few minutes looking at a word cloud.
2. They show where change or action is needed. If people describe customer service as "poor" more often than "good", that is a sign the business should improve it.
3. They make comparisons easy. Two word clouds of customer service reviews for two shops, say in London and New York, show the most common feedback for each and which has the better reviews, or whether they are much the same.
4. They show changes in trends or patterns over time, for example customer service feedback from 2023, 2024 and 2025 side by side.
5. They help people digest lots of information clearly. Some people find a large data table confusing or overwhelming, and visual learners may find this kind of pictographic chart easier to understand.

Finding messages versus conveying them

Word clouds have two key uses. The first is to help find and understand the messages in the text. The second is to help convey information to an audience. You can, of course, use one cloud for both.

Finding messages in text

Word frequencies are a useful summary when the question is very simple. If people are asked for the one word that best describes something, a summary of the frequencies is a sufficient analysis. An analysis of the former UK Prime Minister, Liz Truss, is an example: two separate word clouds were enough to identify that in September the Government was shown as Determined and Strong, and by the start of October the key words were Incompetent, Useless and Untrustworthy.

In general, though, a word cloud is a starting point for the analysis rather than a complete solution: it can speed the work up, but in many cases you still need to read the actual text, as the word counts alone are unlikely to be enough.

Conveying information to an audience

Whether or not a word cloud helped you find the meaning in the text, it can help you show that meaning to others. For this job you can bring in what you know from elsewhere, use colour to highlight associations and change the layout, so the cloud tells the story rather than just the counts.

A worked example: the 2022 State of the Union speech

Ray Poynter walked through both uses in January 2023 with the text of President Biden's 2022 State of the Union speech.

The first cloud counted single words. The more often a word appeared in the speech, the larger it was drawn, so the cloud was led by "america", "people", "year" and "americans". Words such as "a", "the" and "and" were suppressed; words left out like this are known as stop words. America and Americans topping the list is hardly surprising in a State of the Union speech. Ukraine, Russia, Putin, Justice, Covid-19, Health and Economy pointed to key topics.

"Tonight", by contrast, was a narrative device: "so tonight I'm offering a unity agenda for the nation" and "but tonight I say that we will never just accept living with covid-19". For this speech it might make sense to treat "tonight" as a stop word too; How to remove words from a word cloud shows how.

The second cloud favoured phrases over single words, using the scoring WordCloud+ calls Combination Counting, now the default for new projects. It was led by "america", "american rescue plan", "americans" and "cut the cost", and the phrases gave a clearer picture of key topics such as Electric Vehicles, Fight Inflation and Child Care. Word clouds with phrases explains how that scoring works.

It was still necessary to dig deeper. Going back to the original text for the phrase "burn pits" found "breathing in toxic smoke from burn pits", "these burn pits that incinerate waste", "just yards from burn pits the size of football fields" and "cancer from prolonged exposure to burn pits". In WordCloud+, clicking a phrase and opening References shows every passage it came from.

The third cloud was made to convey the message. Drawing on other information about President Biden's plans and a more detailed reading of the text, Ray used colour to highlight associations, with phrases such as "health care", "american rescue plan" and "jobs" in red, and changed the layout of the items. The result shows that the speech covered a wide range of topics, such as Ukraine and the burn pits, and also how central the American Rescue Plan was to President Biden. For the following year's speech, see the 2023 State of the Union contrasted with 2022.

What word clouds cannot do

On its own, a word cloud does only one thing: it illustrates the frequencies of words in the text. When a question asks people for the one word that describes the customer service they have received, the word counts can readily convey the main message.

In most cases, a word cloud is at most a starting point in the search for meaning. If the meaning of the text is closely linked to how often terms occur, a word cloud gives you a start by identifying the phrases that occur most often. If you are looking for a deeper meaning, such as people's underlying beliefs and motivations, a word cloud is unlikely to help. To borrow an analogy from art: we cannot understand da Vinci's Mona Lisa by counting its paint strokes, or how often blue appears rather than green.

Tag clouds and Wordle: a short history

Writing in January 2023, Ray Poynter noted that word clouds had first appeared about twenty years earlier, and that their popularity grew rapidly when Wordle was launched. For many years Wordle (wordle.net; the word cloud tool, not the word game) was the go-to tool for creating word clouds, but by January 2023 it no longer appeared to be available. Wikipedia's entry on tag clouds covers more of that history, including Wordle.

Make your own word cloud

Try it on your own text: paste the text, see the preview, and create a free account when you want to edit and export. Premium adds watermark-free downloads and coding, with a 30-day free trial; see the pricing section. To see how our co-founders have used WordCloud+ for market research and insight, visit our YouTube channel.