This week Twitter added about 70 new locations for “trending topics”, i.e. what Twitter defines as topics, hashtags or key words that are currently getting a surge in mentions. Sweden is one of the new countries for which Twitter displays trending topics and while the service has been up and running only a day or two, we can already start to suspect that it’s accuracy is not entirely perfect yet.
My first point: local trending topics seem to count words that are mentioned in URLs’, giving major media outlets too much impact on the results, since their content is widely shared across Twitter. First evidence, yesterday the tabloid “Aftonbladet” was one of ten trending topics but many of the tweets that showed up, were in fact not mentioning “Aftonbladet” other than the URL Aftonbladet.se, often hidden in a bit.ly-link. Second evidence, today “Expressen” the rivalling tabloid is a trending topic, with the same boost in mentions from URLs as Aftonbladet. I don’t think that this is intentional and I suspect that Twitter would adjust their algorithm if they were aware of the problem.
My second point: the hashtag #cheerupjustin is barely being used at all in Sweden, still it is trending here. The term “English” is also trending, and although a Twitter search in Swedish gives us many results, few of them are in Swedish and/or by Swedish users. It is probably difficult for Twitter to determine which tweets are related to Sweden and therefor tweets from other countries affect our results.
It’s great to be able to track trends also in the Swedish market, but use the trending topics with caution as they may not always be entirely accurate.
I must be getting old because I am frequently baffled by Twitter’s trending topics. If the Sweden is anything like the States then the trending topics will be comprised of 60% teenage popstars, 30% nonsense and 10% actual live breaking news topics of interest.