Home

The year in review, and What's to come

Naturally in this day and age, the path to understanding the past and future goes through twitter.

Now here is the challenge, in order to make this more exciting: What we really need is a statistical analysis of the #code2009 and #code2010 streams (and in particular, their differences). The goal is to post code that does this analysis in the most elegant and succinct way; naturally using the langues du jour earns bonus points.

By Ehud Lamm at 2009-12-31 23:50 | Fun | other blogs | 8628 reads

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Live stat

Live stat pie chart of #2009. Not mine.

By bashyal at Fri, 2010-01-01 00:53 | login or register to post comments

Nice. Of course, we want a

Nice. Of course, we want a lot more, e.g., are there trends over time (are early twitters early adopters?)

By Ehud Lamm at Fri, 2010-01-01 01:24 | login or register to post comments

Twitter

I must say, this whole trend of open Internet diaries (Twitter, Facebook, etc.) strikes me as belonging right up there with the Darwin Awards.

I'm just waiting for some powerful analysis engines to start combining and predicting what these individuals are doing, and keeping its own open diary (and search engine...). If a thief wants to know which houses on his block will be empty of people, just ask! People are tweeting about their family experiences in another state, right now.

By dmbarbour at Fri, 2010-01-01 04:10 | login or register to post comments

Breaking into empty houses

Breaking into empty houses based on twitter information already happened of course. But until this is powered by a program written in a domain specific language I fail to see the relevance of this to LtU...

By Ehud Lamm at Fri, 2010-01-01 05:01 | login or register to post comments

Granted.

I don't really see what popularity analysis of languages volunteered by a self-selected population of Twitter fans has to do with LtU, either. But I would like to see some development of languages to support mapping and mining of plain text (especially across languages and resources).

By dmbarbour at Fri, 2010-01-01 05:30 | login or register to post comments

That's a legitimate

That's a legitimate concern... Of course, I mentioned this for its anecdotal value not as scientific evidence, but I think a discussion of whether this kind of data teaches us anything that is of interest would certainly be on topic... Happy New Year!

By Ehud Lamm at Fri, 2010-01-01 05:49 | login or register to post comments

At different ends, PADS/ML

At different ends, PADS/ML and Hadoop are almost exactly for this ;-)

By lmeyerov at Sat, 2010-01-02 01:14 | login or register to post comments

;-)

Phishing languages will become the huge hit of 2010. It'll be like the create-your-own-punter AOL thing. Ehud will be asked to appear in a commercial endorsing it.

That's my predictions for 2010.

By Z-Bo at Thu, 2010-01-07 17:30 | login or register to post comments