Friday 13 June 2008

Jonathan Feinberg has made a cool application: Wordle. You give it a corpus of text, and it creates a word cloud with size based on frequency. Here are the titles of all my blog posts over the years:


Wordle automatically nestles all the words together, and when you create one, you can set parameters like font and layout algorithm. Very cool.

The thing that surprised me about my title cloud is that it looks like a tag cloud, but these are actually words pulled from the titles, not the names of tags. Maybe all that manual tagging isn't worth it after all?

tagged: » 11 reactions


Josh 9:35 AM on 13 Jun 2008

First time I tried it I got some Python error code. Second and subsequent tries have resulted in "Error: Not Found
The requested URL /create was not found on this server." Bummer, this looks fun!

Ned Batchelder 10:12 AM on 13 Jun 2008

I had some problems with Google App Engine claiming the app was over quota, but then a reload would get me the page, and entering direct text complained about Unicode while uploading a file always seemed to work.

Jonathan Feinberg 10:48 AM on 13 Jun 2008

I apologize for the growing pains, which--I believe--have been corrected. Please try it again, if you have a moment.

David Phillips 10:37 AM on 14 Jun 2008

It's lovely. I just ran article I'm in the middle of writing through it and I could tell right away that it had captured the sense of the article. Both the key technical terms and the some important non-technical words made it through.
Since Jonathan Feinberg has posted just above, let me say it's beautiful. I've never tried to write something like this but, I guess the words are the easy part and the balance is the hard part. The fonts are great and the balance of the sizes and colors are great, too.
I hope you can keep wordle online. I'm thinking of trying to write a little applescript to call wordle from emacs before I close a latex file.

sanatizi 4:56 AM on 16 Jun 2008

kültür, sanat, müzik, resim, fotoğraf, el sanatları, tarih, edebiyat, dünya, insan

Amit C 1:53 AM on 21 Jun 2008

Hi Ned,

Thanks for pointing to it.

Did you try to get it with the non-breaking space working?


nicki 5:48 PM on 22 Jun 2008

I'm having trouble with the non-breaking space working and I can't e-mail Jonathon. Can someone help?

Jonathan Feinberg 5:53 PM on 22 Jun 2008


When you say you can't e-mail me, what do you mean?

Amit C 6:01 PM on 22 Jun 2008

Hi Jonathan,

Pretty nifty tool, thanks for creating it.

Can you please put up, maybe in the FAQ, an example of the non-breaking space for any OS, any browser encoding to be used and one or two fonts it works with.

On Linux(Ubuntu & RHEL 5), I tried few variations including \u00A0, with quotes, escaping, even copying from Character Map and chaning browser encoding (UTF 8n, 16, etc), none of these worked to get words like "bay area" to be considered together.


nicki 10:21 PM on 22 Jun 2008

I click on the link to contact you, a box pops up and it says: could not perform this operation because the default mail client is not properly installed.

Add a comment:

Ignore this:
not displayed and no spam.
Leave this empty:
not searched.
Name and either email or www are required.
Don't put anything here:
Leave this empty:
URLs auto-link and some tags are allowed: <a><b><i><p><br><pre>.