I am fascinated by the conversations formed in the comment spaces around articles and blogs. With the [...]]]> How do you capture the essence of a conversation online? How can we visualize chats and forums? How can user generated content that usually follows news articles and blog posts take a more substantial role in the online conversation?

I am fascinated by the conversations formed in the comment spaces around articles and blogs. With the current way that comments are displayed as long lists, it is virtually impossible to comprehend the broad range of perspectives. For the past few weeks I’ve been looking at the comment space around articles on the US presidential campaign. The LA Times chose to present the thousands of comments (specifically the 7,972 between September 28th until today!). For this project I chose to use java with processing. The goal is to create a clear narrative/s around the essence of these comments, binding them together on different threads. It is not a simple task, as the comments usually do not relate to one another. Moreover, users submit comments usually under fake names – and when you do not know who exactly is talking, it is extremely difficult to have a conversation. So I chose to concentrate less on the ‘who’ and more on ‘what’: what are people saying about the candidates. Which words are they using more, and how they are choosing to describe the candidates.

playing around with the interface and aesthetics

An inspiration to this project is Ben Rubin and Mark Hansen’s Listening Post – displaying real-time conversations from online chat rooms.

(listening post photos courtesy of fenchurch @ flickr)

More updates and a link to the web-app VERY soon

[tags] visualization, campaign, 2008, primaries [/tags]

They intend to understand natural language queries inputed by users on their mobile phones, and to deliver highly relevant answers. The service allows you to retrieve search [...]]]> “Will it rain today?” Apparently mInfo, a Shanghai-based mobile search startup, is working on a service that will be able to answer such a question on our phones.

They intend to understand natural language queries inputed by users on their mobile phones, and to deliver highly relevant answers. The service allows you to retrieve search results using SMS or WAP. Their revenue model is that of a common search service – generating revenues from target advertising and allowing clients to bid for search result ranking. I just wonder how well that works with SMS, since the length of each message is extremely short. I also hope this service has extremely good NLP (natural Language Processing) analysis, since we are ever more sensitive to span in our mobile phones. Somehow it is bearable in our email, but when it comes to phones, and users paying per message they receive, spam messages will not be tolerated.

(source: shanghaimomo)

This same company has recently acquired Guanxi, a city guide which allows you to SMS keywords in English, Mandarin, or Pin Yin from your mobile phone to search locations in cities right across China. The service responds to your SMS with the result/s of your search in the language that you used to search. If there is more than one matched venue for your search criteria, you will be sent a list of the matched venues, replying with the number of the corresponding venue from the match list will get you the address and directions of your selected venue. To get these details in Chinese simply reply with a C.

The service owns two short codes, one for China Unicom users:

and the other for China Mobile users:

Upon signing up to this service, you have to insert your China ID Card Number or your Passport Number in order to receive a one month free service. And that’s the part where myself and probably 90% of the other potential users click the x on the top right of our browser. Interesting service, especially for gwailos with minimal Chinese language skills. But still doesn’t seem ready for mass use.