Students’ and graduates’ mobility is an interesting topic of discussion especially for the Italian education system and universities. The main reasons for migration and for the so called brain drain, can be found in the socio-economic context and in the famous North–South divide. Measuring mobility and understanding its dynamic over time and space are not trivial tasks. Most of the studies in the related literature focus on the determinants of such phenomenon, in this paper, instead, combining tools coming from graph theory and Topological Data Analysis we propose a new measure for the attitude to mobility. Each mobility trajectory is represented by a graph and the importance of the features constituting the graph are evaluated over time using persistence diagrams. The attitude to mobility of the students is then ranked computing the distance between the individual persistence diagram and the theoretical persistence diagram of the stayer student. The new approach is used for evaluating the mobility of the students that in 2008 enrolled in an Italian university. The relation between attitude to mobility and the main socio-demographic variables is investigated.

In such a global economic crisis context, our aims are describing the mobil- ity of the Palermitan graduates in the labour market, identifying the variables that influence most their occupational status and finally outlining a transition probability structure among the states: Work, Search for a Job, Study, Other. The availability of a large amount of longitudinal data provided by the surveys carried out by STELLA (Statistics about Graduates and Labour Market) allowed us to analyze the mobility of the graduates of the University of Palermo among the different occupational states in three different times. We analyze data coming from a disproportionate stratified sample of graduates in 2009, interviewed three different times: one year (2010), three years (2012), five years (2014) after the graduation. To achieve our aim, first we provide a brief descriptive analysis of the main characteristics of the graduates gathered by the three different surveys; secondly we fit a time in- homogeneous multi-state Markov model with piecewise constant intensities; eventually implications from the main results are discussed.