Twitterbot predicts what books you want for Christmas
3 December 2018 | 0
A project by the Adapt cente for research into digital content has found a novel way to come up with recommendations for reading matter over the Christmas period: by scouring your Twitter feed.
By tweeting #SCICOM18, the bot @ReadMeLikeABook delivers personalised book recommendations.
The AI-powered system makes book recommendations based on a topic choice or information in a Twitter feed. If no topic is included, the bot uses the account holder’s Twitter bio as a personalised basis on which to recommend a book. If that yields nothing, it looks for suitable topics in the user’s recent tweets. It remembers past recommendations so that repeated requests give different suggestions.
The bot, created by researcher Dr Tony Veale, has been built with a humour and creates self-deprecating comments. It can even provide a personality analysis of a user’s recent tweets if the user directly responds to #ReadMeLikeABook. A visual representation is attached to the response tweet of the user with a breakdown of personality traits and an associated weighting.
The bot was developed to mark this year’s Sci:Com conference for the people who shape the story of science in Ireland. This year’s keynote will be delivered by actor and science communicator Alan Alda (The Aviator, ER, The West Wing, M*A*S*H). The conference takes place on 12 December at the Ballsbridge Hotel in Dublin.