The introduction of the Internet-of-things (IoT) paradigm allows the use of heterogeneous sensors to monitor vast areas at a low cost. The scalability of the approach depends on the possibility of processing data at a high rate: a wide network of heterogeneous devices provides sensing and status information that aligns to the speed, volume, and variety typical of Big Data systems. One of the fields of application of IoT is the protection of cultural heritage: controlled spaces (eg, museums) can be monitored with a sensing infrastructure that can be easily installed, monitored, and extended, and data can be easily processed in real time to implement alarms or control of environmental parameters such as temperature and humidity or even to perform a continuous analysis of the state of perishable assets. In these cases, if measurements indicate that parameter values are outside the specified range, access to the rooms may be limited or delayed. Of course, IoT is not the only applicable technology for these situations, but it is probably the most flexible state-of-the-art solution. When the space considered is a wide and inhabited area, such as a town, and its cultural heritage is to be protected, scenarios grow in diversity and complexity and human behavior can hardly be ignored. In protected historical areas such as Pompei, Venice, or the center of Rome, in which the vastness of the area, the nature of the access paths and the presence of inhabitants prevent or obstruct any kind of access control, then, as documented by news reports, the probability of site damage or theft is significantly higher than 0. In such scenarios there is also a more complex problem: emergency management. In case of incidents (terrorist attacks, natural disasters, public demonstrations, or rallies), there is no direct escape path planning that can be applied, and the intervention of police, firemen, national guard, or medical aid personnel is hard because of the complexity of the terrain and the behavior of the crowd (eg, panic). Festivals or events are the ideal scenario to plan a terrorist attack or the defacement of a cultural heritage site. Such situations require additional monitoring that adapts to the size of the crowds, the nature of the event, and the alarm level. In this paper, we present a quantitative modeling approach for the planning and management of events in vast, populated cultural heritage sites that relies on adaptive IoT-based monitoring and situation assessment systems in which crowd behavior is included. This approach allows both for the modeling the possible scenarios and the design of the main parameters of the needed computing infrastructure. The main advantage of our approach is its capability of catching, by means of an analytical solution, the dynamics of a high number of model elements with a compact, parameterized description and a low computational cost, with an accurateness that is increasing with the number of agents in the system. The paper is organized as follows: Section 2 details the motivation for this work; Section 3 presents related work; Section 4 describes the modeling approach; Section 5 shows a case study; and Section 6 closes the paper