The compiling of large-scale building inventory is a fundamental step in the evaluation of seismic risk in a region of interest. For classification of buildings in vulnerability classes, different levels of detail could be employed depending on the model adopted for the vulnerability assessment, moving from simplest vulnerability models that identifies classes based on the sole material of load resisting system to more sophisticated ones that requires further specifical information on buildings, such as the construction age, the number of storeys as well as the type of horizontal system in masonry buildings. Generally, the key structural characteristics of exposed building needed for the application of more refined vulnerability models are not easily available. Census data, the most easily accessible source of information on buildings and population in many countries, only provides limited information about buildings features (e.g., storey number, construction age, construction material). With the scope to improve the information quality relevant to exposure, the interview-based form Cartis was recently implemented in Italy by Civil Protection Department within “Territorial Themes” ReLUIS project. Being based on an interview protocol, this form allows to rapidly detect relevant buildings data at urban level: the prevalent building typologies within sub-municipal areas, denominated Town Compartments, are identified and for each of them the percentages of occurrence of significant building’s features for describing their seismic vulnerability (e.g., the layout of vertical structures and the type of horizontal structures for masonry buildings) are detected. The Cartis approach is an extremely useful tool for the compiling of regional scale inventories; however, nowadays the database only partially covers the Italian territory, counting about 390 Italian municipalities investigated until now. This study aims to exploit the actual data available in the Cartis database to build an exposure model that could be applied also in towns where the Cartis form is not compiled yet. To this end, a machine learning approach can be applied. Cartis data available for several municipalities in Campania region is used together with basic information on buildings and population provided by Italian census data (ISTAT) to train a supervised learning algorithm for identification of homogenous urban sectors in terms of buildings features (i.e., Town Compartments) in municipal area. Once the algorithm is implemented, it may be employed in other municipalities to delimit relative Town Compartments and to identify most common building typologies within them. This procedure may allow to compile regional scale inventories and to improve large scale risk analysis thanks to the adoption of a more refined exposure modelling based on Cartis. Moreover, as the interview with a local expert with a deep knowledge of the construction features in the area remains the more reliable approach for the compilation of the Cartis form, the proposed approach may be a useful tool to support and validate interview outcomes, allowing to reduce the compiling phase time, as well.