Simpy Chichimichi

Last week I posted part 1 of the Simpy tagging study, where I looked at how much we tag. That study showed that Simpy users are hard-core taggers, tagging nearly 99% of all links. We also learned that almost all tags are single word tags, with only 2% of all tags being multi-word tags.

While most Social Bookmarking services, including del.icio.us¹, force their users to expose all their links to the public, Simpy lets users mark links as private. No, that does not mean that Simpy is anti-Web 2.0 and openness, it just means I want to give people the control over their own data. The natural questions to ask are: How many people use these features? How many links are actually private in Simpy? And the answer is in the following figure.

Apparently, about 7.3% of all links in Simpy are private links. Apparently, some non-negligible portion of Simpy members do like to keep their links to themselves, and I bet they are not all p0rn links.

Once I found out that a little less than one tenth of all links are private, I started wondering whether there is any difference in tagging behaviour when people tag their private links vs. their public links. After all, they must have some good reasons to keep certain links private, so maybe this affects how they tag them. After a bit of investigation, this is what I came up with:

As you can see, there really is some difference in tagging behaviour when people tag their private tags. Public links have, on average, 4.94 tags associated with them. I believe this is a pretty high number of tags per link, and supports my earlier statement that Simpy users are quite hard-core taggers and are building a very rich Folksonomy. I would love to see numbers from other services. Anyone? But what about tags for private links? Indeed, private links carry even more tags - 6.63 on average. Can we draw any conclusions from this? If we assume that the number of tags associated with a link indicates the importance of the link, we could say that private links are, on average, more valuable to people, as people put more effort into tagging them by coming up with more tags.

So much for private vs. public analysis. If you have some other ideas or requests for additional types of analysis, feel free to email.

In the next study I'll look at chronology of different aspects of activities found in a Web 2.0 Folksonomy service such as Simpy.

[1] I'm sure Joshua & Co. are working hard on private links.