Session 14 - Data & Opinion Mining

Abstract: Modern communication is ubiquitous thanks to the advances in mobile devices and the spread of social networks across the world. It allows people to share ideas, opinions, and experiences. With an ever grown number of users, third party entities promote their services through online advertisements aimed to attract more customers. These services can be even illegal, making this kind of advertisements deceptive. In this work, we study the use of tweeter for human trafficking criminal network. Mining data from tweets, especially to identify hidden information is a complex process, which even requires human intervention. Thus, more sophisticated natural language methods are necessary. In our experiments, the term "punters" has the clearest message related to human trafficking services, we find some phrases that depending on the context could lead to criminal activities, and we show that temporal analysis seems an interesting research area to identify behavioral patterns which certainly can help to combat this kind of criminal networks.

Authors: Danilo Burbano and Myriam Hernández-Alvarez

Abstract: Imagine a dystopia where sensitive, personal information about an individual can be found out by asking an official who works in a public office. The requester does not even have to identify himself to obtain information about the taxes others have paid. After a few direct questions, he figures out national identification numbers, home phone numbers, employer's names, and addresses. The requester can realize that someone is an elder adult, married to a much younger woman, and recently sued for drunk driving. It is not that personal records are being hacked, but that the State is openly releasing this information in a fierce attempt to honor the right of the rest of people to know. 
Sadly, this scenario is becoming closer to reality in Ecuador. Personal information about citizens are not available by request in physical offices of public institutions. Yet, it is worse. Sensitive information of public employees is being released online under the Ecuadorian transparency law. Also, several personal attributes of citizens are being disclosed through web interfaces inspired by a carelessly implemented e-government policy. In this paper, we analyze how these hyper-transparency practices may lead our society to a digital version of the dystopia described above and how to protect citizens from such privacy risks.

Authors: Ana Rodríguez-Hoyos, José Estrada-Jiménez, Luis Urquiza-Aguiar, Javier Parra-Arnau and Jordi Forne