Follow Me On Twitter

Sunday, November 04, 2007

A Quick Jaunt Into Data Mining on Twitter

Holy Twitter, Batman!  It's Boris Groys!

So just who the hell is Boris Groys and why might any of us who Twitter care about him?  Other than the fact that he might be a teacher of art history and philosophy in Germany, or a professor of Slavic studies in New York, what other purposes might Boris Groys serve us?

"Boris Groys is a philosopher, essayist, art critic, media theorist, and an internationally acclaimed expert on late-Soviet postmodern art and literature, as well as on the Russian avant-garde. Dr. Groys’s writing engages the wildly disparate traditions of French post structuralism and modern Russian philosophy."

For those of you who recall the movie, "A Beautiful Mind", you'll recall John Nash was a schizophrenic who was able to discern all kinds of complex patterns in what most people would perceive as garbled gibberish.

twitter keyword activity explored at twitterversetwitter keyword activity explored at twitterverse
Wait, that's not Nash.That was Maximus!


Enter Twitterverse... and Mr Nash is nowhere to be found.

Twitterverse is one of the many cool little play things that Twitterholics can find and play with at the Twitter Fan Wiki.

And that, ultimately, led me to Professor Groys...

twitter keyword activity explored at twitterverse
(Do ya really need me to highlight it in yellow for ya?  Didn't think so...)

So what's the big fuss about a bunch of words surrounding Boris Groys (except that one is a swear)?

Well, the cool thing about Twitterverse is that it allows you to see the words about which people are twittering AND to what degree.  The larger words are words that are being mentioned more often.  As I discovered though, a larger word doesn't necessarily mean a specific number of instances.  As far as I can tell, the word sizes only represent how many times the words are said in relation to each other.  So if all other words were mentioned once, a large word could be one that was mentioned, hypothetically, a mere 5 times.  But there are ways to determine that a little better (see below).

And Twitterverse also allows you to take 1, 5 and 10 hour snapshots which was pretty cool because it got me wondering why Boris Groys was so popular in the last hour and not the last 5 hours.  I'll show you why and how I found out.


TwitterSearch is another cool app found via the Twitter Fan Wiki page. It allows you to search all instances of a given keyword phrase on Twitter's Public Timeline.  The public time line is where all unprotected twitting gets dumped in real time.  So if you go to TwitterSearch and type in something like, oh say, BORISGROYS, you get something like this...

research keyword popularity on twitter with twittersearch
(there were about 10 more by the same guy over a 2hr period)

So our excitement dies down a bit because we see that it's one guy who is enamored of Boris Groys.  In fact, after looking at all of his Twitters, they struck me as having been auto-submitted.  Just like the following...

twitter user keywords explored through twittersearch
(there were a bunch more over a 6 hour period)

That came up during a "two-word" phrase search at Twitterverse resulting in this:

explore behavior patterns on Twitter with Twitterverse
(you can see the same twitterer triggered the term "music feed", as well)

And it was probably done from a feed submission app, ALSO found on the Twitter Fans Wiki page, such as Rss2Twitter:

submit blog feeds to twitter with rss2twitter
(I took the plunge and am auto-submitting my blog feed to Twitter now)

TwitterVision will then allow you to explore a little further to see if a specific term is location-popular.  But wait!  If you go directly to the homepage, you'll get the absolutely addictive world view showing from WHERE all the twitters from the public timeline are coming.  That can be interesting; and will definitely lead to addiction, but we are looking at it for another reason.

When in TwitterSearch, having searched for the term "BorisGroys" and getting the results shown above, you can click on the picture of any/all twitters and it will show you where they live on the map.  Here is one example:

exploring twitter users keywords and behavior on twittervision
(found by adding the Twitter username to the end of the domain: - looks like Brazil to me!)

Anyways, you see what this is getting at?  If someone were to create a Twitter data mining application that could highlight patterns, siphon out the refuse and forecast trends, I think it could potentially be worth a lot of money.  After all, what Twitter affords us is a HAIR-DOWN look into the minds of a great concentration of some very successful about as close to "real time" as possible.

Sure we could throw some big bucks at people/companies who already do some very good research by other means but this is free, and it's right now.  And I can think of a lot worse ways to be spending your time.... like TWITTERING!

Now get to work! ;-)

Make sure to read Part 2:
Further Exploitation via Data Mining on Twitter

Sam Freedom"s Internet Marketing Controversy Blog

AddThis Social Bookmark ButtonAddThis Social Bookmark ButtonHome


Dan Anton said...

Interesting, I didn't even know about Twittering till I read your post. Since you think those were automated will it take away from the purpose of it? or you just showing it's potential for future use. The idea of the words getting bigger when used more is a pretty cool way to represent it. Gonna have to twitter something and see if I get addicted ;)

Sam Freedom said...

At first, Twitter seemed like no big deal to me. I had to wonder if certain big name marketers who were using it were just showing their "childish" side, playing with a kiddie app - and quite honestly, there is a LOT of gibberish, idiotic comments coming from people whom you'd expect to be mature.

But there are also a lot of POWER USERS out there, too. And if you know how to pique peoples' curiousity then to the victor goes the spoils.

I mean, there ARE plenty of intelligent people out there Twittering but if some of them are content to perch on it all day and twitter like a lark, then why not throw them a juicy worm or two and invite them to power up something over at your place for a while? ;-)

As for the automated part, I can't be 1000% sure that's what it was but I'd still wager it. He has 25,000 updates and while I'm still not sure what he's up to, it all does get indexed. Even if the backlinks are crap PR1 or 2 backlinks, 1000s of them WILL add up and WILL point back to his Twitter profile and history.

I'm thinking there's something quite possibly ingenious going on and how interesting it is that my innocent little peek into the Twitterverse brought it up.

I'll certainly let you know more as I discover it.

ps. Check out the 3rd party apps page. Find some marketer whom you know is respected and start following all of his friends and some of them will follow you back.

magazine said...

I really don't get the twitter movement. Is it just me, but what is the point?

web development Dubai said...

This is a nice start, but I suspect this graph is based on a very small sample of the Twitter social network. I spent last week crawling Twitter to create a graph of the social network, and I what I got is very different. I started off graphing friends, but it was so dense that I moved on to mutual friends who have both posted within a given time period (e.g., a day), and it's still so
dense that it's almost impossible to tease apart. The only interesting thing to report so far is clusters by language. The biggest cluster is English and the second biggest is Japanese. Actually, something else that is very interesting is that ranking Twits by number of mutual friends gives a completely different top Twits list than ranking by simply total # of friends.

HP0-Y20 dumps said...

This really is an informative post man. I have got a lot from it. To me the great feature of Twitterverse is that it permits you to look at the words regarding which users are twittering along with to what degree. Thanks for sharing such an informative post.

online internet marketing services said...

Honestly, i never tried Twitter before and as what i have read in this post, it's very interesting.

total12 said...

As far as I can tell, the word sizes only represent how many times the words are said in relation to each other. Buy YouTube Comments

Anonymous said...

I don't cerebrate some of websites offer this typewrite of assemblage.

Anonymous said...

So if all other words were mentioned once, a large word could be one that was mentioned, hypothetically, a mere 5 times.

Jarking said...

The lineament of your articles and listing is large.wp themes

jacking said...

Your blogs and every other content is so entertaining and useful It makes me come back again.
seo services stoke on trent