Global Navigation Satellite Systems (GNSS) are widely used to document the on- and off-site trajectories of construction equipment. Before analyzing the collected data for better understanding and improving construction operations, the data need to be freed from outliers. Eliminating outliers is challenging. While manually identifying outliers is a time-consuming and error-prone process, automatic filtering is exposed to false positives errors, which can lead to eliminating accurate trajectory segments. This paper addresses this issue by proposing a hybrid filtering method, which integrates experts’ decisions. The decisions are operationalized as parameters to search for next outliers and are based on visualization of sensor readings and the human-generated notes that describe specifics of the construction project. A specialized open-source software prototype was developed and applied by the authors to illustrate the proposed approach. The software was utilized to filter outliers in sensor readings collected during earthmoving and asphalt paving projects that involved five different types of common construction equipment.