“Social Media and Public Administration. How the constant monitoring and content analysis of social media can help the PA to improve their online communication“
My Ph.D. scholarship is funded by the INPS (National Institute for Social Security), which is also one of the few Italian PA having an efficient team that is responsible for administering the Twitter profile of the Institute.
From this premise my PhD thesis wants to consider how the PA can improve their communication towards the citizen monitoring and analyzing the online flow of communication. The case study chosen and analyzed is the one of Taranto’s Ilva steel company whose executives (on July 26, 2012) received by the judiciary eight notices of detention and the seizure of the plant. The accusation is environmental disaster.
Before July 2012, the majority of Italian citizens were not aware of the precarious working conditions and the damage that Ilva produced in the city of Taranto. Issues brought to the forefront of public opinion by the extensive coverage given by the media to the story from that point onwards. Issues that affect strongly the INPS as it has to cope with the costs associated with the supply of: 1) pensions for infirmities and disease, and 2) unemployment benefits.
Among the main objectives of my thesis there is an analysis of the communication flows occurred on Twitter and concerning Taranto’s Ilva (hashtag #ilva) since 2011. A first analysis of the data reveals how many people were asking visibility to the major national media with respect to this matter prior to July 2012. Visibility, however, obtained only after the intervention of the judiciary. In second instance building strategies of web content analysis usable by the PA to anticipate future issues and raise awareness of the citizen / worker on matters related to risks in a working environment.
During my university studies (and now my Ph.D.), I developed a strong attraction for the content analysis. This discipline has led me over time to use and develop web applications that are able to collect and later analyze large amounts of textual data.
For those who deal with statistical analysis of textual data, the assets of text generated daily by Social Network Sites such as Twitter and Facebook is a valuable resource but potentially evanescent and often constrained by restrictive Terms of Service.
For all these reasons I started years ago to study a high-level programming language such as Python and became interested in Machine Learning algorithms. Thanks to these tools I was able to understand how to monitor most of the Social Media and I learned the design principles of a Sentiment Analysis (Opinion Mining) model. Trying to do a proper content analysis also means to pay particular attention in the development of analysis strategies not fully automated like, in example, the Sentiment Analysis, and that’s why already during the drafting of my MA thesis I have dedicated myself to the study of CATA (Computer-Assisted Text Analysis) software such as T-Lab and TaLTaC2, that gave me advanced control over the lexical analysis feasible through them. The experience gained in the use of software and collecting textual databases from the web has allowed me to investigate different Italian relevant topics such as the rise of Beppe Grillo’s Five Star Political Movement, the twitter discussion around the Italian Elections of 2013 and the Taranto’s ILVA environmental disaster.