Twitter search helps scientists plot language differences worldwide

With thousands of tweets being set every second, scientists have tapped into the social network as a great research tool.

Before your “Facebook feed alarm” starts sounding, rest assured the researchers just looked at tweets that had been sent in Spanish and used GPS data included in the tweets to plot them on a map.

They were keen to find out how different words for the same things – think pavement and sidewalk in English – are used worldwide.

Is ‘Twitter’ the same in other languages? (Chris Ison/PA)

Using this method researchers Bruno Goncalves and David Sanchez found that Spanish is split into two superdialects.

First up is an urban speech used across major American and Spanish cities. Second is a diverse form that encompasses rural areas and smaller towns and which can be further sub-divided.

They picked out words which are used differently across the Spanish-speaking world – for instance “computer”, which might be referred to as computadora, ordenador and computador; and “car”, which could be auto, carro, coche, concho and movi.

This allowed the researchers to “easily define regional dialects characterised by the set of dominant words”.

Spanish was chosen over English and Mandarin for the study (Manu Fernandez/AP/PA)

The urban dialect was also found in areas where Spanish isn’t the first language – notably certain US states, plus Western European and Brazilian cities.

The full report “Crowdsourcing Dialect Characterisation Through Twitter” can be found on Cornell University’s arXiv website.