Abstract: In one embodiment, a computing system may access data samples associated with a geographic area of interest covered by a communication network. The data samples may be aggregated into a plurality of data points. The computing system may partition the plurality of data points into a first set of data points and a second set of data points using a first threshold of a first network metric. The computing system may determine a predicted gain of the second network metric for a network enhancement operation. The predicted gain of the second network metric may be determined based on a comparison between a first weighted average of the first set of data points and a second weighted average of the first set of data points and the second set of data points. The computing system may generate one or more network optimization recommendations for the geographic area of interest based at least in part on the predicted gain of the second network metric caused by the network enhancement operation.

Abstract: In one embodiment, a computing system accesses a plurality of quality of experience metrics from a plurality of geographic areas, respectively. The system identifies a first geographic area of the plurality of geographic areas connected by a first communication network that has low quality of experiences based on quality of experience metrics and a set of criteria. The set of criteria comprises one or more thresholds corresponding to the quality of experience metrics. The system determines, for the first geographic area, a severity of the low quality of experiences. The severity is based on a connection demand metric associated with each geographic area. The system identifies a second communication network that has one or more quality of experience metrics higher than the first communication network. The system sends, to the client system, instructions for automatically changing from the first communication network to the second communication network.

Abstract: In one embodiment, a computing system may receive a request for an optimization recommendation of a geographic area of interest covered by a communication network. The computing system may determine a network traffic trend associated with the geographic area of interest based on a current number of data samples that may be aggregated into a plurality of data points. The computing system may predict a value of a number of data samples for a future time associated with the geographic area of interest, based on the determined network traffic trend and the current number of data samples. The computing system may predict, based on the determined network traffic trend and the predicted value of the number of data samples at the future time, network traffic associated with the geographic area of interest at the future time, and send instructions for presenting the optimization recommendation based on the predicted network traffic.

Abstract: In one embodiment, a computing system receives a request for a quality of experience report associated with a plurality of geographic areas connected by a communication network. The system accesses a plurality of quality of experience metrics from the plurality of geographic areas. The system identifies one or more of the plurality of geographic areas that have low quality of experiences based on their respective quality of experience metrics and a set of criteria which includes one or more thresholds corresponding to the quality of experience metrics. The system determines for each identified geographic area, a severity of the low quality of experiences. The severity is based on a connection demand metric associated with each geographic area. The system sends instructions for presenting the quality of experience report which indicates the severity of the low quality of experiences for each identified geographic area.

Abstract: In one embodiment, a computing system may receive a request for a measurement project of a communication network for a geographic region of interest. The computing system may identify network metrics for control regions. The control regions may comprise geographic regions connected by the communication network. The network metrics may comprise features that measure a change in a value of a corresponding network metric during a particular time period. The computing system may determine, based on a comparison between a first control region and a second control region, whether a first network metric has a change greater than a pre-determined threshold after a first network deployment. The computing system may send, responsive to the request for the measurement project, instructions for presenting a measurement result summary for a second network deployment based on the determined change after the first network deployment is greater than the pre-determined threshold.

Abstract: In one embodiment, the system identifies one or more geographic areas covered by a communication network. The system determines, for each identified geographic area, a congestion metric for the identified geographic area based at least on a difference between a first and second reference point on a network speed curve, wherein the network speed curve represents download speeds for a volume of traffic in the identified geographic area. The system identifies one or more network traffic congestions in one or more of the identified geographic areas based on a comparison of the respective congestion metrics of the identified geographic areas to a threshold congestion metric. The system sends, to one or more operators of the communication network, one or more alerts about the identified network traffic congestions.

Abstract: In one embodiment, a computing system may access data samples associated with an access point of a communication network. The data samples may be collected at an application level from client devices associated with the access point and aggregated into data points correlating a first network metric to a traffic-volume metric. The system may partition the data points into a first and second set of data points using a first network metric threshold. The system may determine trends of the traffic-volume metric with respect to the first network metric based on regression analysis on the first and second sets of data points. The system may predict, based on the trends of the traffic-volume metric with respect to the first network metric, an amount of time until a network-capacity metric of the access point meets a traffic-loss threshold. The network-capacity metric may be determined based on the traffic-volume metric.

Abstract: In one embodiment, a computing system may determine network metrics for geographic regions connected by a communication network. The geographic regions may include a geographic region of interest impacted by a network deployment. The system may identify candidate control regions by clustering geographic regions of into a first group based on metric feature of network metric shared by the geographic regions. The system may select control regions by identifying a second group of geographic regions from the first group of geographic regions based on a comparison between the region of interest and each of the first group of geographic regions. The system may compare a first network metric of the geographic region of interest to first a metric baseline determined based on the one or more control regions. The system may determine whether the first network metric has a change greater than a threshold after the network deployment.

Abstract: In one embodiment, a computing system may access data samples associated with an access point of a communication network. The data samples may be collected at an application level from client devices associated with the access point and aggregated into data points correlating a first network metric to a traffic-volume metric. The system may partition the data points into a first and second set of data points using a first network metric threshold. The system may determine trends of the traffic-volume metric with respect to the first network metric based on regression analysis on the first and second sets of data points. The system may predict, based on the trends of the traffic-volume metric with respect to the first network metric, an amount of time until a network-capacity metric of the access point meets a traffic-loss threshold. The network-capacity metric may be determined based on the traffic-volume metric.

Abstract: In one embodiment, a computing system may receive a request for an optimization recommendation of a geographic area of interest covered by a communication network. The computing system may determine a network traffic trend associated with the geographic area of interest based on a current number of data samples that may be aggregated into a plurality of data points. The computing system may predict a value of a number of data samples for a future time associated with the geographic area of interest, based on the determined network traffic trend and the current number of data samples. The computing system may predict, based on the determined network traffic trend and the predicted value of the number of data samples at the future time, network traffic associated with the geographic area of interest at the future time, and send instructions for presenting the optimization recommendation based on the predicted network traffic.

Abstract: In one embodiment, a computing system may collect data samples associated with a geographic area of interest covered by a communication network. The system may aggregate the data samples into data points. The system may split the aggregated data points into two sets of data points using a first threshold of a first network metric. The system may determine a trend of a second network metric over the first network metric based on regression on the two sets of data points. The system may determine a predicted gain of the second network metric for a network enhancement operation based on the trend of the second network metric and a reference value of the first network metric. The system may generate network optimization recommendations for the geographic area of interest based at least in part on the predicted gain of the second network metric.

Abstract: In one embodiment, a computing system receives a request for a quality of experience report associated with a plurality of geographic areas connected by a communication network. The system accesses a plurality of quality of experience metrics from the plurality of geographic areas. The system identifies one or more of the plurality of geographic areas that have low quality of experiences based on their respective quality of experience metrics and a set of criteria which includes one or more thresholds corresponding to the quality of experience metrics. The system determines for each identified geographic area, a severity of the low quality of experiences. The severity is based on a connection demand metric associated with each geographic area. The system sends instructions for presenting the quality of experience report which indicates the severity of the low quality of experiences for each identified geographic area.

Abstract: In one embodiment, the system identifies one or more geographic areas covered by a communication network. The system determines, for each identified geographic area, a congestion metric for the identified geographic area based at least on a difference between a first and second reference point on a network speed curve, wherein the network speed curve represents download speeds for a volume of traffic in the identified geographic area. The system identifies one or more network traffic congestions in one or more of the identified geographic areas based on a comparison of the respective congestion metrics of the identified geographic areas to a threshold congestion metric. The system sends, to one or more operators of the communication network, one or more alerts about the identified network traffic congestions.

Abstract: In one embodiment, a computing system may collect data samples associated with a geographic area of interest covered by a communication network. The system may aggregate the data samples into data points. The system may split the aggregated data points into two sets of data points using a first threshold of a first network metric. The system may determine a trend of a second network metric over the first network metric based on regression on the two sets of data points. The system may determine a predicted gain of the second network metric for a network enhancement operation based on the trend of the second network metric and a reference value of the first network metric. The system may generate network optimization recommendations for the geographic area of interest based at least in part on the predicted gain of the second network metric.

Abstract: In one embodiment, a computing system may determine network metrics for geographic regions connected by a communication network. The geographic regions may include a geographic region of interest impacted by a network deployment. The system may identify candidate control regions by clustering geographic regions of into a first group based on metric feature of network metric shared by the geographic regions. The system may select control regions by identifying a second group of geographic regions from the first group of geographic regions based on a comparison between the region of interest and each of the first group of geographic regions. The system may compare a first network metric of the geographic region of interest to first a metric baseline determined based on the one or more control regions. The system may determine whether the first network metric has a change greater than a threshold after the network deployment.

Abstract: In one embodiment, a computing system may receive user inputs from an interactive user interface specifying a communication network and a geographic region of interest. The user interface may retrieve measurements of the communication network and a network deployment taken on the communication network. The system may generate project for measuring impact of the network deployment on the geographic region of interest. The system may access network performance data associated with time periods before and after the network deployment. The system may determine whether the impact is greater than a threshold based on incremental network metrics which may be determined based on the accessed network performance data and using one or more measurement methodologies. The system may send instructions for presenting a visualization of the impact within the user interface. The visualization may describe the geographic region of interest and incremental metrics with respect to the geographic region of interest.

Abstract: In one embodiment, a computing system determines quality of experience metrics and root-cause metrics for geographic areas connected by a communication network based on application data. The system determines that one or more of the geographic areas have low quality of experience based on the quality of experience metrics and a first set of criteria which includes one or more first thresholds for the quality of experience metrics. The system determines causes for the low quality of experience in each associated geographic area having low quality of experience based on the root-cause metrics and a second set of criteria which includes one or more second thresholds and is looser than the first set of criteria. The system determines a solution recommendation for the low quality of experience in each associated geographic area based on the causes of the low quality of experience of that geographic area.

Abstract: In one embodiment, the system identifies geographic areas covered by a communication network. The system determines, for each identified geographic area, a network performance metric for the identified geographic area based on a difference between: (1) a first average network speed of the communication network in the identified geographic area during prior time periods in which the communication network is busy, and (2) a second average network speed of the communication network in the identified geographic area during second prior time periods in which the communication network is not busy. The system compares the respective performance metrics of the geographic areas to a threshold network performance metric, which is determined by a congestion-analysis machine learning (ML) model. The system identifies traffic congestions in one or more of the identified geographic areas having a determined network performance metric below the threshold network performance metric.

Abstract: In one embodiment, the system identifies geographic areas covered by a communication network. The system determines, for each identified geographic area, a network performance metric for the identified geographic area based on a difference between: (1) a first average network speed of the communication network in the identified geographic area during prior time periods in which the communication network is busy, and (2) a second average network speed of the communication network in the identified geographic area during second prior time periods in which the communication network is not busy. The system compares the respective performance metrics of the geographic areas to a threshold network performance metric, which is determined by a congestion-analysis machine learning (ML) model. The system identifies traffic congestions in one or more of the identified geographic areas having a determined network performance metric below the threshold network performance metric.

Abstract: In one embodiment, a computing system determines quality of experience metrics and root-cause metrics for geographic areas connected by a communication network based on application data. The system determines that one or more of the geographic areas have low quality of experience based on the quality of experience metrics and a first set of criteria which includes one or more first thresholds for the quality of experience metrics. The system determines causes for the low quality of experience in each associated geographic area having low quality of experience based on the root-cause metrics and a second set of criteria which includes one or more second thresholds and is looser than the first set of criteria. The system determines a solution recommendation for the low quality of experience in each associated geographic area based on the causes of the low quality of experience of that geographic area.