A movement analytics platform can generate instructions for collecting video data from a site that includes a roadway. The instructions can be provided to a device having augmented reality capabilities, wherein the instructions include content displayed by the device to indicate an area at the site that an operator of the device is to position within a field of view of a camera. A data feed received from the device can include video data corresponding to the area and contextual data to annotate the video data based on observations by the device operator. The data feed can be processed to derive movement analytics associated with the area at the site (e.g., classifications, locations, speeds, travel directions, and/or the like for one or more objects depicted in the video data). The device can be provided with additional augmented reality content based on the movement analytics.