Museum preferences analysis: an item response model applied to network data

Firenzecard is the official museum pass of the municipality of Florence that allow the visit of more than eighty collections and exhibitions located in Florence and in the surrounding area. Firenzecard provides a huge amount of information concerning the paths of visits followed by tourists as well as some individual characteristics (e.g., country of origin). In this contribution we focus on the data relating to the 127,092 cards sold in the year 2018 which correspond to a total of 884,389 visits to museums. First, we use the instruments proper of the (social) network analysis (Kolaczyk, 2009) to provide a description of the relations among the museums in terms of tourists’ preferences. Second, we estimate a binary latent class item response model (Bartolucci, 2007) to detect unobservable (latent) classes of tourists that are homogenous with respect to their propensity to visit museums. In particular, this analysis is aimed at identifying museums whose attractiveness differs among latent classes of tourists