Abstract: Techniques are described to detect and classify, with high accuracy, events associated with objects using image data analysis and computer vision. In an embodiment, an image capturing device captures an image of an area, which includes objects such as fixed objects. The process determines that an event associated with an object occurred by providing the captured image as input data to an ML model. The ML model is trained to identify such objects and events associated with the objects. The process determines image coordinate information of the event identified within the image and, based at least on mapping data of the area to the image coordinate information, determines that a fixed object of the area is associated with the event. Metrics may be calculated for the fixed object and detected events.

Abstract: An analytics provider receives a request for supplemental content to be displayed to a user by a user device. The request includes a network identifier identifying the user device on an access network over which the user device receives content. The analytics provider transmits the request for supplemental content and the network identifier to an analytics module provided by the analytics provider and executing at a network operator that manages the access network. The analytics module collects data describing behavior of the user device on the access network, determines at least one learned usage characteristic based on the data describing the behavior of the user device on the access network, selects a supplemental content item based on a target usage characteristic associated with the selected supplemental content item and the at least one learned usage characteristic, and provides the selected supplemental content item for display on the user device.

Abstract: Techniques are described to detect and classify, with high accuracy, events associated with objects using image data analysis and computer vision. In an embodiment, an image capturing device captures an image of an area, which includes objects such as fixed objects. The process determines that an event associated with an object occurred by providing the captured image as input data to an ML model. The ML model is trained to identify such objects and events associated with the objects. The process determines image coordinate information of the event identified within the image and, based at least on mapping data of the area to the image coordinate information, determines that a fixed object of the area is associated with the event. Metrics may be calculated for the fixed object and detected events.

Abstract: An analytics provider receives a request for supplemental content to be displayed to a user by a user device. The request includes a network identifier identifying the user device on an access network over which the user device receives content. The analytics provider transmits the request for supplemental content and the network identifier to an analytics module provided by the analytics provider and executing at a network operator that manages the access network. The analytics module collects data describing behavior of the user device on the access network, determines at least one learned usage characteristic based on the data describing the behavior of the user device on the access network, selects a supplemental content item based on a target usage characteristic associated with the selected supplemental content item and the at least one learned usage characteristic, and provides the selected supplemental content item for display on the user device.