If you’ve ever digitally hailed a ride, the chances are high that you’ve used Lyft. As one of the world’s leading ride-sharing organizations, Lyft is valued at over $11 billion and has become an integral part of many commuters’ lives.
A key to Lyft’s success is its easy-to-use mobile application. The health and growth of Lyft’s business depend on the superior functionality of its mobile app which both riders and drivers use at different points and for different purposes—before, during, and after a ride.
Lyft collects millions of data points and metrics from its application operating on multiple timescales: seconds when matching passengers and drivers; minutes when calculating pricing; hours when measuring market levers; and days or weeks for budgeting.
But with multiple geographic dimensions on top of that, Lyft simply didn’t have the resources to manually monitor every metric it gathered to detect anomalies in its data.
To accurately detect anomalies at scale that could signal larger problems and require immediate attention, Lyft turned to the automation and machine learning (ML) capabilities of AWS and Anodot, an AWS Partner Network (APN) Advanced Technology Partner and AWS ML Competency Partner.
Anodot’s AI-powered time series analytics solution, built on AWS, uses advanced machine-learning algorithms to overcome limitations that humans bring to manual data analysis, identifying potential problems in real time without having to inspect multiple dashboards manually. Using Anodot and AWS services such as AWS Kinesis Streams, Lyft is able to detect business incidents quickly to understand the scope and root causes of issues across millions of metrics.
Anodot is an Artificial Intelligence analytics solution that discovers revenue leaks and brand-damaging incidents in real time. Its automated machine learning algorithms continuously analyze business data, detect the business incidents that matter, and identify why they are happening by correlating across multiple data sources.