Update: I’ve stopped active development on FeedVis. It was a learning project, and after spending some time trying to do some refactoring, I’ve decided that it’s just too ugly to fix. You’re still welcome to download the code and email me with minor problems, but I won’t be fixing anything significant from here out.
FeedVis is an interactive tagcloud for groups of rss feeds. This is the development page; here’s the demo page.
I liked the idea of making a word cloud to examine a group of feeds; however, though there were several options, I wanted more. A plain tagcloud from a feed is static: all the words happen now. Conversations change, and I wanted to see that. I wanted more context, too–information about who was using the words, and how. Hence, FeedVis. It lets you compare tagclouds from various authors and times, as well as giving you access to the posts that actually use the words.
Download
If you’d like to check out the code behind FeedVis or run it on your own server, help yourself. There’s no database to set up, so installation is generally pretty easy, although some folks have had some problems. If you run into trouble, feel free to drop me a line.
download feedvis 1.03.zip (requires at least php 5.2)
Reviews
Several fine Internet publications have posted about FeedVis:
ReadWriteWeb
Information Aesthetics
OPML News
RSS4Lib
Data Mining: Text Mining, Visualization and Social Media
Knowbodies
e-clippings
TextmodeInfobib (deutsche)
Outils Froids (français)
Genbeta (español )
Topify
Write to Reply
CogDogBlog
technologyscan.com
Thanks to
Chirag Mehta for the idea of combining timeline and tagcloud, Richard Heyes for his php port of the Porter Stemming Algorithm, John Resig and the developers of the amazing jQuery library, Paul Bakaus and Brandon Aaron for their jQuery dropshadow plugin, Jorn Zaefferer for his jQuery tooltip plugin, and Ariel Flesler for the jQuery.ScrollTo plugin. And if I left anyone out, let me know.
19 Comments
This looks cool! But I have problem: I’d like to use it with a list of Norwegian blogs, but all I get is a list of short, common words, the equivalents of this, that, is, where and so on… I havn’t looked at the code yet, but could there be some way of making a list of stop-words, words that do not show up in the visualisation?
Thanks for your interest, Magnus. Yes, there’s no problem in adding a list of Norwegian stopwords to the others (I already have French, Spanish, English and German). In fact, I’ve already addedthis list. However, looking at the list I’m still guessing that we’re letting some common words slip through. If you send me a list of any common words that I haven’t included, I’d be glad to add it.
There are, however, two more significant problems: First, as you can see, all of the words are displayed in simple ASCII encoding, meaning that dicritics don’t appear. Second, the stemming algorithm I’m using works only for English words. So, while FeedVis knows that “school” and “schooling” are the same word, it won’t be able to make a similar judgment in other languages. Both problems result from me not anticipating the amount of interest this project would get; it was originally just a learning project for me.
Both problems are fixable. I’ve got my graduate school applications due in a week, so I won’t be able to look at FeedVis till after then. I’ll try to update it before the holidays, though.
That does look like a very swish way of keeping up to date with trends in news and could be extremely useful. I hope the applications go well.
I wish there was a way of integrating this with a Google Reader, because this is SO AMAZING.
I will take a look at it (at the code)… maybe I can figure a way of making it work better with Portuguese…
It doesn’t know synonyms in Portuguese, does it?
I’ve just wrote a post about your great tool on my blog Outils Froids, a french blog for knowledge workers.
It would be again greater if it takes accented words. Actually it just cut them.
Feedvis is really a great tool and I don’t understand why there is no more buzz about it. Hope my post help : http://bit.ly/2g3C
@Cindy: Sorry for not getting your comment up sooner; I’ve been getting over a cold. Thanks for the kind words. No, it doesn’t know synonyms in Portuguese, but it doesn’t really know synonyms in any language. Perhaps you’re speaking of stemming, where words with the same stem (like learn, learning, and learned) are reduced to one. Here, it’s just English, I’m afraid.
If you’re comfortable working with PHP, it should be pretty easy to get a Portuguese stemmer working, though; just replace the stemmer.php class with a Portuguese version and you’re all set.
Also, I know my list of Portuguese stopwords is pretty incomplete; feel free to send any more and I’ll include them.
@Christophe: sorry about the delay; been feeling under the weather. Thanks for the review on Outils Froids; my French is not so great, but what I can understand of your blog, I like. I agree with you that one of the key challenges of the next few year is going to be helping user of the web filter and condense the vast amount of information available.
I’m sorry that FeedVis handles accented words so badly; it’s a side project that I don’t have loads of time for right now, but I am planning some improvements over the next month or so. I’ve gotten a very positive response from the blogosphere en Français, so I’d really like to improve the way FeedVis handles French in particular. Feel free to send me any other suggestions.
Must admit, when I first looked at this I didn’t think it worked so well, but when I found a project I needed it for and looked closer - it looks awesome! thanks!
Bjr,
Super outil mais non prise en compte de l’accentuation de la langue française…
Merci.
Do you have any more info on installing it Jason? I’m keen to have a play, but not sure how to get it running on a web server. I’m still a little new to all this stuff. Any advice for the newbie? Would it work on an offline setup via WAMP at all?
Hi. It’s wonderful tool and application.Thanks to author.
I would like use it with russian blogs. What about to add russian stopword list?
I can provide subj.
Thanks.
Denis
@derf: Sorry that FeedVis doesn’t handle accented words so well. It’s on the to-do list.
@Dean: There is some information on installation in the text file that comes with the zip package. If you try that and still have trouble, feel free to shoot me an email. If your server has no web access, there really wouldn’t be much point, since it wouldn’t be able to access any information about blogs.
@Dennis: To include your list of Russian stopwords, just add them to stopwords.txt if you’re running FeedVis on your own server.
Just wanted to say thank you for this. It’s a fantastic tool.
Hi, Can you offer more information on how to trigger the feedreader with a cron job. I’ve been looking at the files and running them manually but nothing seem to make any difference. I know how to set up a cron job but maybe I’m not getting the command right. Thanks.
I can get the PHP script to run. It shows the graph with the posts per day, but when I select a particular date to see the tag cloud, the animation runs, but does not actually display the cloud.
Have you seen this behavior before? Any thoughts?
Here’s an update: The javascript error logger in FF is kicking back this:
Error: cal.win is undefined
Source File: frontend/js/tagcloud.js
Line: 178
I’m not a programmer, so no idea how to debug, but thought I’d pass along in case other people were seeing this problem. And any advice on how to solve would be greatly appreciated!
-R
Sure enough, when I comment out that line, it works fine. Gonna leave it commented for now, but would love to get that fixed so I could actually see the dates when I hover over the graph. Thanks!
Joss, I never really set up v2 to run on a cron job; there’s no script to update everything at once. Instead, users update their own accounts when they visit them.
I’ll see about adding an update_all file in the next couple days.
@Rick: I’m not sure what’s causing that, as I can’t seem to replicate the problem. What browser are you using? Which version of FeedVis?
2 Trackbacks
FeedVis RSS Feed Tag Cloud Generator…
FeedVis [jasonpriem.com] is an online tag cloud generator with some additional interactive features. Users can select specific time periods, common blog themes or individual blog feeds. Individual tags can be further explored to read specific blog pos…
[...] source code is available for download if you want to run your own memetracker (kind of). Via [...]