Web

Analytics

Trend spotting is something I’ve often heard researchers say they want to do. In fact, if you’re a futurist or a practitioner of the Lead User method, then trend spotting is a necessary part of what you do.

I’ve got a few ideas for how to spot trends that leverage social media.

One idea is based on a linguistic pattern I’ve noticed when people discuss trends. Here are some examples:

“Keeping distance between the car ahead of you and coasting as much as possible. Some people take it to the extreme and they call it Hypermiling.” (source)

They call it ‘sexting’: teenagers using cell phones to send nude or semi-nude pictures to a girlfriend or boyfriend.” (source)

They call it exergaming or exertainment – a combination of video games and physical activity.” (source)

Do you see the pattern? When a trend becomes significant enough, people begin to name the phenomenon, something I wrote about a while back in my post on Challenges for Natural Language Processing. So if we analyze the language for patterns such as “they call it ‘x'”, it might be a good way to identify trends.

Another pattern to look for might be, “more and more people today are …”. That would help identify trends that haven’t become significant enough yet to be named. In some cases, those might be the trends you’re more interested in anyway. After all, you may be the one who’s trying to spot the trend and name it yourself — you trend spotter, you. ;-)

Another technique I’ve considered is that of monitoring Wikipedia for new entries that are suddenly being frequently edited. I believe
Wikipedia has an edit log that could be monitored.

What techniques have you seen for trend spotting?