In this study, we presented a novel approach that combines expert knowledge with machine learning to map open spaces in mountain regions. Rather than relying exclusively on physical attributes, we used a Delphi survey to reach a consensus among experts on what constitutes open spaces, considering both physical aspects and their subjective perceptions of the landscape. Using machine learning methods, we extrapolated this information and created a map of open spaces in the evaluated mountain regions. In addition to providing a validated and detailed map, this approach fostered collaborative decision making and facilitated processes of knowledge redefinition, ultimately leading to improved and shared outcomes. This dataset contains all the scripts as well as the final map, but not the starting raw data, as these are either already provided elsewhere or not openly accessible.