<p>Integrating traditional statistical sources on the labour market with online data availability is challenging as, together with official statistics, online information offers detailed and timely information about the specific skills required by businesses. In the light of this, online job advisements (OJAs) data carried out by Lightcast for Italy for ICT professionals, from the year 2019 to 2022, are used to analyse the evolution of skills required, paying particular attention to skill type and regional detail. From a methodological point of view, a latent class transition analysis is applied. Skills are classified by using as indicator variables the importance of each skill in order to: (i) classify skills into homogenous classes characterised by different levels of importance, (ii) assess whether changes in latent class membership occurred over time, and (iii) evaluate the effect of skill type and region where job is required on the latent status membership and transition probability. Empirical findings uncovered three latent statuses of skills importance from the low to the high intensity. The levels of skill importance remain quite stable over the period considered, but some interesting changes are detected for some skill types and some Italian regions.</p>