Social Media Data per lo studio della disoccupazione giovanile italiana: il progetto LikeYouth

[Autom. eng. transl.] In the years of crisis, the subject of work has acquired a growing attention in Italian families. It is a transversal concern, but one that particularly affects the new generations. The employment rate of young people already before the crisis had fallen below 35%, only to sink in the worst years of the recession to around 20%. During 2016, it slightly rose again, but remained abundantly below 25%. The objective of the LikeYouth project is to describe the characteristics of NEETs (Not in Education, Employment or Training) in terms of conditions and attitudes - useful for cognitive purposes on the phenomenon but also potentially able to provide help indications for interception and engagement measures - going beyond not only the data of public statistics, but also experimenting the combination with innovative sources such as social media data. In particular, the objectives of the study are many: Automatically identify the Neet population through their digital traces (likes on Facebook pages) to reach them easily for high-performance interventions; Discover behavioral and individual characteristics and personality of the population for a better understanding of the phenomenon, integrating the results with traditional surveys; Predict potential individuals at risk to create privileged communication channels, and promote targeted policy measures and campaigns. To achieve the objectives set, LikeYouth (www.likeyouth.org), a Facebook App in Italian and English, was created in collaboration with the Department of Statistical Sciences of the Catholic University of Milan, ISI Foundation, Cariplo Foundation and Istituto Toniolo. It is an innovative data collection tool that aims at understanding young people's behavior and profiling by integrating digital information and self-reported assessment from a questionnaire. It is an open, extensible, modular and adaptable system for questionnaire administration and integration with social data. The project is based, in detail, on recent literature results, which show a strong predictive relationship emerged among the typical manifestations of interest of Facebook (the so-called "likes") and some traits and personal profiles. The sample consists of 9358 individuals (Youth Report Survey), aged between 18 and 33 years. About 20% (exactly 1858 subjects) of the young people to whom the application was proposed made full access to LikeYouth. The comparison between the sample total and the sample of subjects who accessed LikeYouth showed no significant differences with respect to the most important socio-demographic variables. The two samples were also compared for some relevant psychometric indicators and for use of technology, without highlighting any statistically significant difference. For these reasons the sample of young people who accessed LikeYouth can be considered representative for all purposes of the Italian youth population and shows how the use of digital tools (often cheaper for the client and practical and appealing for the participants) can guarantee complete scientific attention to a large-scale survey. The purpose of this preliminary and exploratory study was to do prediction and clustering, as well as an association study. We tried to predict the working condition (by dichotomizing it for simplicity in two categories, Neet and Not Neet) on the basis of the digital behavior that act as predictors, we tried to interpret and classify the subjects based on their tendencies and behaviors and to investigate possible associations between the different conditions and digital behaviors put in place, for particular cohorts of the population of relevant interest.