Vast Active Library and Information Science blog. From a recent library science graduate in Wellington, New Zealand. A focus on reference and current awareness tools and issues, especially free, web-based resources.

This page is powered by Blogger. Isn't yours?
Friday, December 30, 2005
Opinmind: Positive and Negative Blog Views  
Seen on Micro Persuasion and LibraryStuff, reviews of OpinMind, a blog search engine that attempts to measure bloggers' opinions of various topics, organisations, or people. For example, type in "Bill Gates" and you get two columns of results, one positive and one negative - "I love what Bill Gates does" vs "I hate Bill Gates". Results can be sorted by date or by relevance, and the "sentimeter" measures the percentages of positive vs negative comment.

I like this tool a lot, it's something of a shame that it "only" searches 1.7 million blogs (Technorati indexes 21 million or so). Inevitably, given the nature of language, it isn't perfect - for example, if someone writes about wearing an 'I love Bill Gates' shirt in order to annoy someone else, that would count as a positive result.

Also, Opinmind seems to search a lot of journal type blogs, Livejournal and Xanga, which lessens its impact - I'd rather hear what a techie blogger has to say on a given subject than what a bored teenager has to say, though there is some value in the latter's opinions.

Stephen Cohen noticed some flaws, pointing out that the results for 'librarian' and 'librarians' were almost opposite each other in terms of positive vs negative comments.

What I found more interesting was using 'library' as a search term. The results were generally positive (68% to 32%). It's the 32% that interests me. These are customers (or potential customers) of the library, and they're saying what they don't like about it. Isn't this great information? It's like free market research! I also tried searching on the names of specific libraries, but didn't get many hits (there were a few for the really big libraries - New York Library, British Library - but that's it).

What would be great is if there was some way to extract information from a blog and determine which library [or whatever] the blogger was referring to. For example, my profile says that I'm in Wellington. If I post about the public library, it's fairly obvious to a human that I'm referring to Wellington City Libraries. Wouldn't it be great if Opinmind could understand this, so that I could search Opinmind for "Wellington library" and return one of my posts that mentioned the library, even if it didn't specifically mention Wellington?


Comments: Post a Comment