Predictive Analytics for Network Data

A major telecommunications company needed to improve the performance of their networks to meet the requirements of their Service Level Agreement. The volume of data on their network was enormous and changing constantly, with weather, season, device traffic, and the continuing technology changes all affecting the network’s performance. The telecom could see that their traditional process of manual hypothesis testing and statistical analysis could not keep pace with the speed and scale of the data on their networks. They looked for a solution that would automate this process and found Emcien.

Using EmcienPatterns, the telecom’s Center of Excellence team was able to analyze all of these combined data sources for all of the combinations of data points that predict an interruption of voice or data services on one of the country’s largest cellular networks. Their analysis automatically extracted all of the predictive rules for dropped calls, network slowdown, and other degraded service levels. These results identified many opportunities to improve the service levels across the network that were hidden to other analysis tools.

 The team knew that the traffic on their network would shift over time and that even using Emcien’s automated predictive rules analysis by itself would not keep pace with their data. To address this enormous flow of data, the team used Emcien’s APIs to integrate automated predictions into their existing data structure. Connecting these systems, the team was able to automate the entire data workflow, from extracting new data, calculating the predictive rules, to predicting on new data as it is created. The predictive rules are now applied to data in real-time to predict service interruptions, triggering preventative maintenance and avoiding service interruptions.

Business Value:

• Improved service levels ensuring maximum uptime.
• Optimized maintenance with prescriptive scheduling.