A recent article gives details about an online program developed by a team of researchers at the Twente University in Enschede, Netherlands, that can detect a Twitter user’s age and gender. The tool can guess how old a user is and tell men from women by simply tracking the content they post on Twitter, without needing a name or a photo.
The tool analyses the last 200 tweets posted by a user, who were simply required to enter their username in the online program to calculate age and gender. Then it uses the words and phrases from a database containing data from nearly 3,000 Twitter users and compares the contents to see how they correspond with the different ages and genders, Dong Nguyen, a doctoral student in computer sciences at the university, explained.
Generally, the difference between men and women is quite stereotypical, Nguyen said. Women like to talk about their nails, while men talk about football, she added. As to age, younger Twitter users talk much more about themselves and use emoticons more often, while older users express their thoughts in longer words and sentences.
The program can currently analyse Tweets in Dutch only, but the researchers are working together with the Meertens Dutch language and culture institute to update the tool for other languages. They also plan to expand the program’s capabilities and make it able to decipher posts on other social media sites, such as Facebook.