these series of experiments around ranking give a simple overview of the traffic related to the r-echos articles. It gives insight on trends and zeitgeist.
the main issue with this techniques is that it gives only the most accessed clues, which produce an information-rich model.
What would be the next steps:
1/ the less accessed articles and tags – what is the less trendy? is there a rare pearl in there?
2/ the favorites – there are always a few websites in the whole that interrests me a bit more
3/ the weight of a tag – the more articles are defined by a keyword should affect the whole system
4/ the weight of an article – …
3 and 4, at first sight, are based on an information-rich model as well; however the datas generated by this bias could be used in a different way
update:
solution 3/: (beta) each time an article is defined by a keyword, this keyword get one point. 10 most accessed tags via articles