Our newly releases Web security product, GlobalView URL Filtering, is an API-integrated module that resides in existing Web security offerings to provide advanced URL filtering. The innovative “Data Cloud URL Filtering” architecture enables classifications to be performed “in the cloud.” This structure also moves the database to the cloud, freeing local devices from storage limitations.
This change has a significant impact on the coverage and accuracy of the URL filtering solution. You can read more about it in our latest white paper: “Defending Against Modern Web Threats: Introducing Data Cloud URL Filtering”.
I decided to take our new technology for a test drive, and as a twitter newbie, I thought it would be interesting to check the categorization of URLs on twitter.
How the experiment was conducted
I used the twitt(url)y Web site as a source for the most frequently tweeted URLs. According to their Web site, twit(url)y is a “service for tracking what URLs people are talking about as they talk about them on Twitter.” You can read more about twitt(url)y in their about page.
These are the most discussed URLs on twitter:
I extracted all the URLs from this list into a text file:
I then ran the list through our URL Filtering engine for analysis.
Fair warning, the results do not represent a sample of the entire twitter community, rather just the most tweeted URLs as taken from the twit(url)y Web site.
Here is what I found:
I have to admit – no big surprises here!
News and technology makes up 63% of the subjects users tweet about. Taking into consideration the people I know from twitter, I actually expected it to be higher.
What’s more interesting are the small percentages of the categories that make up the “Other” category: entertainment at 3% and pornography at 1%. The most surprising figure to me was sports at 1%! Come on, I am a fellow geek myself, but only 1%?! What can I tell you? I’m disappointed.
It would be interesting to run a test on a representative sample of URLs in the entire twitter community, and not just the most tweeted. I guess if I run such a test in the future, we will see some spam and malware URLs as well, since I believe that the twitter service, like any Web 2.0 environment, is sometimes exploited for malicious activities.
Maybe in the next experiment…For now, I’ll contribute my share and tweet about sports… the Australian Open is in its final stage! That’s worth tweeting about, isn’t it?