Featured Journalism Notes Projects Social Classification

Automatically generated profiles?

Tracking tags
This image shows the most popular tags connected to the articles I have read today

The coming weeks we will be further updating the news website. The last weeks we have added interesting things on the database level and back-end of the system. Now it is time to bring some of those ideas to the front-end.

One of the things we have done is making tags more important. After using and testing with it we noticed the combined tagging methods we use give a very interesting and relevant database with tags. The tags have more value and semantic relations than I thought they would have. has over 35.000 articles in its database with over a 100.000 tags, 10.000 of those tags are unique.

What we have added is a system that tracks the tags of the articles you read. With this information a metadata profile is created. New articles that enter the site will be matched to your ‘profile’ and if there is a match this will be a recommended article. There is also a tag relation mechanism to create a more semantic relation.

We will do the same for your friends, since news is part of a social experience. We are adding groups as well. If you have online friends using the website, you and your group of friends creates a semantic profile as well. New articles will be recommended that fit to your group of friends.

Your profile
What about this tag based profile based on what your read? The profile could be private or hidden. It could also be open or even exportable. For example could connect to web services and get the highest rated and most recent blog posts about subjects you like and recommend these to you. Or you could connect the information to other profiles to create a more rich or enhanced experience on other platforms as well.

Would you like to take this profile? And can you think of a service that could serve you better when it has a collection of news themes and subjects you like?

Concept design personalized frontapage
The concept design for the frontpage. The page automatically orders the information based on what is most relevant in general and for you.

6 replies on “Automatically generated profiles?”


After a few days of using the beta version of the website I already have 1240 unique tags linked to my profile.

Image posted by

Interesting. for news. Not sure if it works (does one crave related news, similar to craving new music?), but the concept is interesting.

There is some difference. can tell you about old music you might like as well, or music you don’t yet know. can only look at new news coming in and when a new article matches your existing profile it can recommend the article (or make it more important). It doesn’t add news it only makes subtle personal changes to the hierarchy of the presentation.

If all your friends read about technology and beaches, news about technology and beaches will be recommended to your group, since this article is something that connects all your interests and probably creates conversation (or watercooler talk).

If it works well it could go out and find related articles by other sources as well, although this is a step further, since you don’t want to recommend articles that aren’t relevant, or duplicates.

I think there are great services out there aggregating relevant stories that can connect to.

Comments are closed.