I found a nice Python module for language guessing that is based on trigram vectors. The idea is very simple: a document is analysed looking a triplets of characters, the frequency of three character sequences is calculated.  When treated as a vector, this information can be compared to other trigrams, and the difference between them seen as an angle.  The cosine of this angle varies between 1 for complete similarity, and 0 for utter difference.  Since letter combinations are characteristic to a language, this can be used to determine the language of a body of text.

This methods are known as N-gram speech recognition algorithms.

I applied this module into my tagger to make it switch between different language parameter files. You can see the results in the difference of tags that are generated without and with the language guesser module below:

2005-10-03 13:42:01,542 – main – INFO – — Tagging session started

2005-10-03 14:05:08,076 – main – INFO – The number of messages tagged is: 190

2005-10-03 14:05:08,122 – main – INFO – The number of new tags created is: 162

2005-10-03 14:05:08,123 – main – INFO – The number of messages dumped is: 47

2005-10-03 14:05:08,123 – main – INFO – — Tagging session ended

Tags: natural language processing, tagging

The rise of corporately owned malls… private CCTV security systems… radical demographic and cultural diversity… The dominant urban story of the past 20 years has been of creeping privatisation and social fragmentation resulting in the loss of space for shared experiences between people.

The challenge now is to find a new story that reconnects public, private and civic spaces alike, and can provide the impetus for growing the shared life of our towns and cities. This agenda applies as much to an out-of-town shopping centre as it does to a neighbourhood park.

Based on in-depth studies of three British towns and cities Cardiff, Preston and Swindon, People Make Places explores how the best public spaces are created by people and communities themselves. The book sets out the forms of governance, design principles and everyday uses that can help boost people’s participation in public space and the wider public life of their town or city.

Copyright notice: the present content was taken from the following URL, the copyrights are reserved by the respective author/s.

Tags: society, urban exploration, urban planning

Designing websites is different from traditional media such as print and television. In spite of the usability guidelines being fairly well known, some of the common mistakes recur even on popular sites. Designers need to be careful and follow these guidelines to improve a website’s interface.

Web-usability guru Jakob Nielsen speaks with Larry Magid about designing websites that capture and sustain the user’s attention.

Tags: usability

A nice hack to show the path of the hurricane Rita. The colors of the flags are used to show the intensity of the wind. I found this kind of visualization really nice.

Tags: google, hack, information visualization, maps

Sophie is a virtual robot that lives in a Kitchen World, she interacts with people to learn how to bake cakes. At Media Lab, they are doing a research study about how people teach her.  We hope people can have fun teaching this little task to Sophie and we can learn a little about the teaching process along the way!  If you volunteer for our study, you’ll to play the Sophie game online and then fill out a survey.

Meet Sophie here…

http://robotpatrol.media.mit.edu/kitchen/

This game is being used as a research platform to learn a little about how they can allow people easily teach machines new things. In this study, players teach the Sophie robot agent to bake a cake. While watching the agent try to bake the cake on her own, players teach by sending various messages via the mouse.

Their hope is that people can have fun teaching an agent a simple little task, and that they can learn a little about the teaching process along the way.

Tags: constructivism, learning technology, microworlds, new technology

Placeopedia, is a simple site where you can connect Wikipedia articles with places, and then make use of our database either to browse, or syndicate the whole lot. We hope that lots of people will connect places they know with their corresponding Wikipedia article, and then our syndicated data can be used as a general geographic lookup table for Wikipedia.

Tags: google, hack, information visualization, Location Based Services, maps, tagging

I found this last post of Martin Terre Blanche extremely intriguing.

George Siemens is definitely onto something with his ideas about networked learning. His latest contribution is a piece arguing that learning design should be about designing ecosystems rather than about designing courses:

“My latest contribution is a piece arguing that learning design should be about designing ecosystems rather than about designing courses:

“We should be focusing on designing ecologies in which learners can forage for knowledge, information, and derive meaning. What’s the difference between a course and an ecology? A course, as mentioned is static – a frozen representation of knowledge at a certain time. An ecology is dynamic, rich, and continually evolving. The entire system reacts to changes – internal or external. An ecology gives the learner control – allowing her to acquire and explore areas based on self-selected objectives. The designer of the ecology may still include learning objectives, but they will be implicit rather than explicit.”

I definitely subscribe to this idea of rethinking the way we propose courses and we think about teaching. Personally I strive for a middle-earth approach between the open ended, multi-brainset, constructivist approach and the self-regulated, computer assisted collaborative learning. I think the way in should be finding a script that structures the interaction and leaving enough room in that script to support multiple attitudes/abilities.

Tags: constructivism, learning technology, connectivism

A set of tools for field biologists to collect, organize, and share their research. These tools comprise mobile components (e.g., an in-the-field photo annotation tool) and lab components (e.g., the ButterflyNet interface for viewing multiple data sources linked by their metadata, such as time and location).

Tags: ecology, field tools, hack, interaction design, Location Based Services, new technology, mobile learning, tagging

Graffiti Archaeology is the study of graffiti-covered walls as they change over time.  The grafarc.org project is a timelapse collage, made of photos of graffiti taken at the same

location by many different photographers over a span of several years. Most of the photos are from San Francisco, over a timespan from the late 1990’s to the present.

Using the grafarc explorer, the user can visit some classic graffiti spots, see what they looked like in the past, and explore how they have changed over the years. One of the part I like the most is the timeline bar. Something that reminds me so much of the timeline I developed for the Biosphera project. This one is nicer though.

Tags: graffiti, information visualization, interaction design, Location Based Services, maps, psychogeography, urban exploration

Now I understand why sometime Jamie or Fabien are not blogging much.  In fact I spent a couple of days on my Part of Speech Tagger, that should extract relevant keywords from the geolocalised messages left by the users in the STAMPS system.

After a couple of hacks around the POSTagger of the University of Stuttgart, I managed to write my own Python extension (thanks to Shuja and Patrick for the hints).

It doesn’t output much at the moment, or at least the output is not so self evident. However I added some lines in the logger to have some sense of stats:

2005-09-28 18:34:49,625 – main – INFO – — Tagging session started

2005-09-28 18:44:26,705 – main – INFO – The number of messages tagged is: 190

2005-09-28 18:44:26,710 – main – INFO – The number of new tags created is: 282

2005-09-28 18:44:26,710 – main – INFO – The number of messages dumped is: 3

2005-09-28 18:44:26,713 – main – INFO – — Tagging session ended

Tags: hack, information metric, information retrieval, Python