The MIT Technology Review recently reported on Future Gazing with Search Data in an interesting article by Katharine Gammon. She notes that computer scientists have praised Web-search data as a way to spot emerging trends. However, recent research shows that the results are mixed and context dependent.
For example, Erik Brynjolfsson, director of the MIT Center for Digital Business, says that Web search queries provide the most accurate predictions in cases where people do research before they make a purchase. His research has shown that a rise in home sales can be predicted from Web search queries.
On the other hand, researchers at Yahoo analyzed search queries and found that such searches are not always the best way to spot a trend. They looked at the volume of search queries related to a particular movie, song, and video game up to six weeks before each came out. However, the traditional critics were just as good in most cases.
It is hard to determine intent from searches and then there might be spammers in operation that compromise the data. The simple volume of searches is a fairly rough measure. The Yahoo researchers said that search data may be useful when a small improvement in prediction accuracy could have a big impact--for example, in the financial world.
The Darwin Awareness Engine™ takes a different approach to looking for trends that can be complementary to these search results. Instead of looking at the volumes of searches, it looks at the volume of conversations about a topic. The BuzzTape™ reports topics that are receiving rising attention in real time. The Scan Clloud™ looks at relationships between sub-themes and the main topics of interest. Since the content self-organizes itself, it is not possible to compromise the results through SEO or spam.
Comments