摘要： In recent years the availability of mobile indoor way finding technologies, Internet-of-Things sensor infrastructures, and geographic information systems has dramatically increased. These systems have been introduced for many reasons, such as: increasing accessibility to users with reduced mobility, optimizing logistics, or reducing the environmental impact of facilities. While these technologies can bring some new insight in how buildings are used, they all suffer from limitations in coverage, accuracy, or scalability. However, when they are combined, a new window into the understanding of human mobility and building utilization is opened. This paper introduces methods to increase accuracy of data obtained from an off-the-shelf commercial indoor navigation by including Internet of Things (IoT) sensor infrastructure and geographic information through a Geometric Constraint Model (GCM). Furthermore, a Mobility Graph (MG) is introduced to visualize trajectory distributions. A simple experiment was conducted to validate the methods. For this, a long narrow hall was equipped with iBeacon infrastructure, an indoo.rs Navigation instance, and a few cheap Raspberry Pi based sensor stations. The environment was mapped using state-of-art geodetic measurements, and a set of recorded experiments were made. In such a long and narrow environment the indoor navigation system is unable to resolve features along the short traverse axis. Without a combined analysis, a divider in the environment is not discernible in the MG, while results using the GCM introduced in this paper clearly indicate the presence of this divider. Automating this approach, can bring indoor location analytics from a descriptive to a prescriptive regime. Furthermore, the application of MG provide a useful tool to better understand large indoor navigation data sets.