As public institutions, museums have a keen interest in providing their visitors with an experience. To continuously develop exhibitions, museums require both qualitative data from visitor books and reliable quantitative data, such as visitor numbers and paths taken. Currently, these are scarcely collected within museums. With Theia's crowd tracking, we are developing a straightforward system capable of capturing and processing real-time visitor numbers, dwell time, and movement data. The system is based on camera images, which are analyzed using machine learning. Real-time data and analyses are provided to museum staff via a web-based dashboard. With Theia, museums can make their exhibitions more relevant to their audience. For example, highlights of an exhibition can be identified from the visitor's perspective, and then appropriate actions can be taken within the museum.