TagCrowd visualization: State of the Union

As an exercise in visualizing the words that characterize the course of United States history, I used TagCrowd to create two word clouds for the president’s ‘State of the Union Address,’ 2002 vs. 2011.

One is by U.S. President Bush on January 29, 2002. The other is President Obama on January 25, 2011. Click on the thumbnail to see the full cloud.

State of the Union Address word cloud, Bush 2002 vs. Obama 2011
State of the Union Address word cloud, Bush 2002 vs. Obama 2011

New version, features + fixes

A new version of TagCrowd recently went live. There are a few major fixes and some new features:

1. Major overhaul of the web spider for URL/webpage text sources. You’ll find that a lot more sites will work properly now (notably, Wikipedia).

2. You can now get multi-word phrases to stay together in the cloud (e.g. New York). More info here.

3. Major overhaul of the error reporting system. Now if there are problems with your text source or cloud, you’ll see a more informative error. Behind the scenes, I’m now able to better track what errors people are having, and fix them faster.

4. Updated the Help page.

TagCrowd how now been running since Summer of 2006, almost 5 years, and been visited by almost a million different people in that time. It long ago graduated from Beta status, but ‘Beta’ is still part of the logo as a reminder to always improve and better serve the needs of TagCrowd users.

Customized HTML clouds for embedding

It’s now easier than ever to customize the size and look of your TagCrowd text clouds when embedding them in HTML web pages and blog posts.

You’ll now find a new “CUSTOMIZE” section near the top of the HTML Embed code where you can customize some of the CSS styles to suit the style of your webpage. Custom styles include font and font size, overall cloud size, margins, padding, borders and background color.

In the near future we’ll introduce controls for changing the color and fonts without having to edit CSS.

(As always, you can edit the CSS that lies outside the customize section, though it’s advanced and we can’t provide support for that.)

Save clouds as PDF

I’m pleased to release a long-requested feature. You can now save your text clouds in Adobe Acrobat PDF format. That means they are scalable for printing at any size. You’ll get a letter-sized document that you can scale to your own needs.

I’ve also streamlined the interface to make it easy to save your cloud in one of three formats:

  1. HTML embed (easy insert into your webpages)
  2. printable/full screen (good for creating jpg/gif/png images of your clouds for websites and docs)
  3. PDF (ideal for printing large or small, email attachments, reports)

I’m planning on adding the option to create both landscape and portrait PDFs.

As with any new feature, post questions or comments on this blog post.

Raised file size limits

Today we’re releasing a feature that has been highly requested, especially by those who do research and want to do visual analysis of long transcripts, survey responses, etc.

Raised file size limits. After some code wizardry, you can now input very large text files into TagCrowd using the “Upload a File” method. The limit is 6 megabytes — larger than the complete works of Shakespeare.

Go big.

International languages

After a crash course in the intricacies of the Unicode standard, I’ve introduced basic support for international languages in TagCrowd. There’s a new option to select the language for your text. This will remove the common words in that language from your text cloud.

“Basic support” currently means languages based on the Latin alphabet (i.e. most of Europe), and all accented characters are converted to plain characters. Currently supported languages include Czech, Danish, Dutch, English, Finnish, French, German, Hungarian, Italian, Norwegian, Polish, Portuguese, Romanian, Spanish and Swedish.

Since this is the first international version of TagCrowd, there will certainly be some bugs. Please let us know if you find any. And if you want TagCrowd to work with your language, please send us a list of common words in your language.

Printable version

By popular demand, I’ve added a little checkbox in the options to display your word clouds full screen. This should simplify things a bit if you’re taking screenshots or printing.

Web scraping and other new features go live

The most common feature request I used to receive was the ability to input a URL and have TagCrowd retrieve the web page text automatically. Since my development efforts are prioritized primarily by what I receive requests for, that feature went live a few months ago to much grateful thanks. I’m pretty sure it’s how most people are using TagCrowd these days.

Another feature I get a lot of requests for is a way to create clouds that only include words above a certain frequency. As of today, you can do that too.

I’ve also taken the opportunity to clean up the interface a bit, revamp the Stoplist editor and fix some outstanding bugs. As always, keep your feedback coming, good and bad.

Photos from CSCL

TagCrowd demo

Here are some images of TagCrowd being demonstrated at the International Conference on Computer-Supported Collaborative Learning (CSCL) in New Brunswick, NJ. (Click on the photo to see the rest).

I created word cloud stickers for every presenter at the conference (all 147 of them) based on the title and abstract of the article they wrote for the conference. There just happened to be a space on everyone’s conference-provided name badge that was the perfect size for the stickers so it was easy for people to display them. I distributed the stickers throughout the conference and you could see people pointing and referring to them whenever they introduced themselves.

It was fun to instigate this at such a large scale. Next time I’ll try to work with the conference organizers directly instead of trying to distribute them all on my own. I sure met a lot of people that way, though!