Scientists at the University of Portsmouth have found that correctly labelling your tweets can vastly improve your experience on Twitter.
They have also identified how hashtags work and identified the four stages of a twitter event.
A hashtag is a word or a phrase prefixed with the symbol #. Phrases should not have any spaces between the words.
Twitter users add hashtags to their tweets, but many users do not label their tweets with the most relevant hashtags.
Dr Mohamed Gaber
“A hashtag is like a virtual filing system. Users who get their filing in order from the outset will improve their Twitter presence,” said Dr Mohamed Gaber, senior lecturer in the School of Computing, who led the study.
“Strong hashtags mean a user’s views are more likely to be read, as they will be easily searchable, and they will be more likely to be included in interesting debate and the exchange of knowledge.
“This could lead to them being considered an expert on a subject, building trust and improving their reputation.”
Twitter users can click on a word that is highlighted with a hashtag and they will be shown all tweets that have used that hashtag.
The top ten most popular hashtags are called ‘trends’ and are listed on the site’s front page.
This is particularly important in the event of breaking news, as users can search a hashtag to find up-to-the-minute information.
“If we take the recent London 2012 Olympics as an example, the main hashtag would be #olympics, a hashtag that is still active today. Someone may then add the hashtag #London2012 and additionally even #cycling or #parking or any number of other subjects.
“It is important that tweets are layered properly, and that companies establish a strong hashtag for their brand or event from the outset, or they could fall foul of bad filing, losing feedback and customer confidence.
“Hashtags have helped companies anticipate potential bad PR moments. During 2012 iTunes, GAP and MySpace all listened to social media and performed u-turns on new designs as a result of negative public opinion,” Dr Gaber said.
Using a technique called Transaction-based Rule Change Mining, Dr Gaber and his team looked at thousands of tweets relating to how occurrences are talked about on Twitter, from natural disasters such as the Japanese earthquake, to international events such as the London Olympics, and breaking news stories such as the Boston bombings.
The team have also identified the four stages of a Twitter event, from its outset to its conclusion.
The lifespan of a hashtag can be as short as a couple of hours, or can last for years in the case of some slow burning science stories such as news related to the Large Hadron Collider.
“Although there is no time scale to a Twitter event, there are four very clear stages. A hashtag will initially emerge, then it will be taken up by thousands of users, then it will be used to tweet unexpected information, and then eventually it will die.” said Dr Gaber.
The four stages are as follows. Stage one is the ‘new’ hashtag. This phase describes the emergence of the hashtag, which at this point is solely used as a way of passing on information like a natural disaster, a death of a celebrity or the opening of an event.
Stage two is the ‘emerging’ hashtag. This phase determines whether a hashtag has longevity, and is the phase in which a large number of people on Twitter are spreading the news. This hashtag is still the same or very similar to the hashtag used in stage one.
Stage three, the ‘unexpected’ hashtag sees the emergence of unexpected tweets, taking the hashtag in different directions and being combined with different hashtags. Jokes, political platforms or complaints are all examples of this.
Stage four is the ‘death’ of the hashtag. The hashtag, and the conversation, is over. The information is no longer of interest to the Twitter audience, and the hashtag dies.
The team is now looking to apply their findings to case studies of recent events such as the death and funeral of Margaret Thatcher and the recent Boston manhunt.
Dr Gaber’s research will be presented on June 9 2013 at The 12th International Conference on Artificial Intelligence and Soft Computing, Zakopane, Poland by Mariam Adedoyin-olowe, the PhD student at the University of Portsmouth, who works on this project. You can view the paper here http://eprints.port.ac.uk/11158/1/TRCM.pdf
