Gosti, GiorgioMenconero, SofiaAmadei, Chiara FloriseFanini, BrunoCampana, StefanoFerdani, DanieleGraf, HolgerGuidi, GabrieleHegarty, ZackaryPescarin, SofiaRemondino, Fabio2025-09-052025-09-052025978-3-03868-277-6https://doi.org/10.2312/dh.20253099https://diglib.eg.org/handle/10.2312/dh20253099The recent spread of metaverse technologies has propelled the digitalization of cultural heritage, particularly the production of 3D models of historical artifacts and sites. This context presents new challenges and opportunities, among these arises the possibility of investigating the design of public engagement and cultural dissemination initiatives through the analytical study of user behavior. With this objective, we developed Procezo, a modular data analysis suite that facilitates the processing and aggregation of user experience data via an easy-to-use web interface, specifically engineered for cultural heritage applications. Indeed, with immersive XR devices, motion-tracking tools, sensors, and online web applications, we can easily record users' experiences in virtual or real environments. From these recorded experiences, with Procezo's specifically developed web-based analytics, we may obtain crucial insights into user interaction patterns. Procezo is part of a larger pilot developed under the H2IOSC project named ''Interlumo''. The pilot is divided into three stages: data capture (Kapto), data processing (Procezo), and data inspection (Merkhet). These stages are based on a strong modular design, both at the logical and software levels. The logical separation allows the implementation of these stages together or separately, and the software separation allows us to run the stages on separate dedicated servers. This modularity allows for greater reuse and scalability. We demonstrate the application of Procezo in data cleaning and preprocessing protocols, as well as its implementation for machine learning (ML) algorithms for pattern discovery, specifically through kernel density estimation (KDE), a reliable non-parametric density estimation methodology. Our implementation is based on a graphical web interface that allows analysts to share and compare different machine learning (ML) pipelines. The presented suite improves the quality and efficiency of the analysis process and enables collaboration between domain and analytics experts. Under the H2IOSC project, we assess Procezo on visitors' experiences exploring a virtual reproduction of Cerveteri Etruscan Tomb, which were captured during remote public exhibits and dissemination events. This approach can be easily applied to several case studies, ranging from interactive installations, to online applications, with the objective of accelerating the detection of interaction patterns.Attribution 4.0 International LicenseCCS Concepts: Mathematics of computing → Exploratory data analysis; Information systems → Asynchronous editors; Data analytics; Data mining; Human-centered computing → Systems and tools for interaction designMathematics of computing → Exploratory data analysisInformation systems → Asynchronous editorsData analyticsData miningHuman centered computing → Systems and tools for interaction design10.2312/dh.2025309910 pages