Research into Twitter User Location Data - Co-Authored by Pembroke Fellow Prof. Min Chen - Published Online

Research conducted by Professor Min Chen - Pembroke Fellow and Professor of Scientific Visualisation - and his colleagues at the Oxford e-Research Centre was published online this month.

The paper, entitled 'I Know Where You Live: Inferring Details of People's Lives by Visualizing Publicly Shared Location Data', was co-authored by Prof. Chen, Dr Ilaria Liccardi and Dr Alfie Abdul-Rahman.

Its findings suggest that the location stamps on a small number of Twitter posts can be sufficient for an average user to find out another user's home address and place of work.

The researchers first collected real life geo-location data from tweets in Boston, USA, before asking the data owners to specify where they actually were. They then conducted an empirical study involving 45 participants to determine how accurately these users could infer the location of the original data owners.

The results of the study suggested that Twitter users' real-life locations can be inferred with a relatively high accuracy using both visual and textual representations and just a handful of location stamps.

The research paper was presented at the Special Interest Group on Computer-Human Interaction (CHI) 2016 Conference in San Jose, USA, on Monday 9th May.

Prof. Chen commented on some of the most interesting findings of the research: 'On one hand, the study sends a warning about how easily data privacy can be breached. On the other hand, the findings also suggest that human intelligence may be under-utilised in many data intelligence applications. It is perhaps more cost-beneficial for computers to support human intelligence, instead of staking all investment on machine intelligence.'

On the further avenues of exploration in this area, Professor Chen added: 'In parallel with this empirical study, there was a theoretical development at Pembroke, where a theoretical measurement for analysing the cost-benefit ratio of human- and machine-centric processes in data intelligence was formulated.'

The mathematics for the research project was first used when Prof. Amos Golan, from the American University in Washington DC, paid a two-week visit to Oxford in March 2015 and stayed at Pembroke College. A video summary of his joint work with Professor Chen is available on Vimeo.

Stories based on the findings of this research have also been published in media outlets including Science Daily, The Financial Express, and the Daily Mail.