[video] – The Filter Bubble – From Human to Algorithmic Gatekeepers

Web personalization and personalized recommendations are recently gaining more and more interest. Companies like Amazon, Google, Netflix, The New York Times, Facebook, Twitter, … already personalize their products in different ways. If you take Google’s search results as an example. Have you ever noticed that a friend of you gets different search results as you do for the same search query? If you never have noticed just try it out it’s really worth noting. Another example are Amazon’s product recommendations which are for example based on your purchases, your product ratings and so on.

Eli Pariser explains in the following TED Talk how “human information filters” get substituted by algorithmic ones, which means how recommendation engines filter information for you. Have a look at the video is is really worth watching:

Do you know other examples of web personalization or recommendations engines? Please leave me a comment at the end of this post.

Enhanced by Zemanta

[video] – How is Hadoop used at Twitter?

Image representing Hadoop as depicted in Crunc...
Image via CrunchBase

In the following video Dmitriy Ryaboy, a Twitter Analytics Engineer and a former Cloudera Intern, explains how Twitter uses Hadoop and Pig. Enjoy the video and have a good weekend!

[vimeo 11110059]

Enhanced by Zemanta

[video] – HBase and Pig: The Hadoop Ecosystem at Twitter

I have just found this very interesting video dealing with the implementation of HBase and Pig in combination with Hadoop at Twitter:

Grabeeter – Grab and Search your Tweets offline and online

Grabeeter - Grab and Search your Tweets

Grabeeter (@grabeeter) has just been launched and I think it is a very useful tool.

Grabeeter enables you to grab your tweets which means that you are able to store your tweets on your local harddrive in a structured format (xml at the time). Using the Grabeeter Client you are also able to perform searches on your local stored tweets.

This solves in parts the following problem. Maybe some of you have already noticed that if you have written more than 3200 Tweets you are not able to access your first tweets anymore due to Twitter access restrictions.

If you register on Grabeeter before you have reached the 3200 Tweets on Twitter you are able to access all your written tweets in the future. Grabeeter archives your tweets and enables you to export them in a structured format (XML and JSON at the time). You can also use Grabeeter Client to directly access your interesting tweets on Twitter again.

Just go to Grabeeter and register with your twitter username. All the other work is done for you. You can afterwards export and search your tweets online or using Grabeeter Client offline.

For all developers we also provide a small Grabeeter API in order to access your tweets using an application you developed.

Have fun and we are happy to get feedback from you.

Related articles

Twitter’s use of Cassandra, Hadoop, Pig and HBase for highly distributed Data Processing and Analysis

Kevin Weil, Analytics Lead at Twitter recently gave a presentation on Twitter’s use of Cassandra, Pig and HBase. Specially interesting is how Twitter uses Hadoop and Pig in their data analysis process.

[slideshare id=3806196&doc=nosqlattwitter-nosqleu2010-100421124212-phpapp02]

(via @kevinweil)

Another great presentation from Tobias Ivarsson gives an overview on NoSQL:

[slideshare id=3735857&doc=nosqlfordummies-100415085745-phpapp01]

(via @thobe)

Reblog this post [with Zemanta]

What will the web look like in five years?

Eric Schmidt, CEO of Google was answering this question in an interesting interview at Gartner Symposium/ITxpo Orlando 2009. ReadWriteWeb has summarized the highlights of Eric’s interview as follows:

  • Five years from now the internet will be dominated by Chinese-language content.
  • Today’s teenagers are the model of how the web will work in five years – they jump from app to app to app seamlessly.
  • Five years is a factor of ten in Moore’s Law, meaning that computers will be capable of far more by that time than they are today.
  • Within five years there will be broadband well above 100MB in performance – and distribution distinctions between TV, radio and the web will go away.
  • “We’re starting to make significant money off of Youtube”, content will move towards more video.
  • “Real time information is just as valuable as all the other information, we want it included in our search results.”
  • There are many companies beyond Twitter and Facebook doing real time.
  • “We can index real-time info now – but how do we rank it?”
  • It’s because of this fundamental shift towards user-generated information that people will listen more to other people than to traditional sources. Learning how to rank that “is the great challenge of the age.” Schmidt believes Google can solve that problem.

Chinese-language content, no differences between distribution channels and real time information are some very interesting topics which I really have to dive in. If you feel interested then have a look at the video of the interview:

[youtube lHxub_yQfig]

(via ReadWriteWeb)

Reblog this post [with Zemanta]

[video] – Trouble with the Twittersphere

Just found a nice video about twitter and it’s users called the Twittersphere.

[youtube PN2HAroA12w]

(via blogdrauf.de)

Reblog this post [with Zemanta]

[video] – Evan Williams on “How Twitter’s spectacular growth is being driven by unexpected uses”

The next video is about twitter’s growth by the unexpected use of the service. Very intersting video!

This video I found on the gorgeous TED Talks. Enjoy watching.

(via TED)

Reblog this post [with Zemanta]

Scientist developes a Brain-To-Twitter Interface

Image representing Twitter as depicted in Crun...
Image via CrunchBase

Adam Wilson, biomedical engineering doctoral student at University of Wisconsin-Madison has developed a brain to computer interface which makes it possible to directly write tweets through thoughts. ReadWriteWeb writes:

Technically, what Wilson did was come up with an interface combining an Electroencephalogram, or brain wave monitor, with an on screen keyboard for selecting letters. The system lights up each key on the keyboard but is able to notice a difference in brain activity when the desired letter for input is lit. Wilson compares it to clicking through multiple letters when texting on a mobile phone.

Once you’ve found a new way to input text – what are you going to do with it? Use it to Twitter, of course!

Clearly, there’s some gimmickry going on in the news of Wilson’s interface. Who knows if this is better or worse than saying that a technology is developed to assist physically disabled people when it’s really going to be used by the military? Wilson does say that the technology will be helpful for people with active brains but immobile bodies. Now they’ll be able to Twitter, among other things, he says. Fair enough.

[UPDATE] Here is a little video showing how the Brain Twitter Interface workes:

[youtube 205dHV55XWQ]

Personally I think it is a great idea to use the currently hyped microblogging service for research activities.

(via ReadWriteWeb)

Reblog this post [with Zemanta]