Algorithms Developed to Use "Word Bursts" to Identify Trends
Earlier this week New Scientist reported that a researcher at Cornell had developed algorithms to detect word bursts in text documents and that these bursts may help to identify trends or new ideas. The article suggests that the new finding in this research has to do with the rate of increase in mentions of a term.
Although this represents the results of new research, some work has already been done in this area. Google produces a report on search trends that they call Google Zeitgeist. This is a weekly or monthly snapshot of recent search queries, not a time series analysis.
Some webloggers cited the New Scientist article as another justification for Google's acquisition of Pyra, but that seems unlikely for a number of reasons that have been discussed here previously.