Follow Datanami:
June 5, 2018

Waze App Gets Anodot Time-Series Upgrade

Anodot, which focuses on using machine learning techniques to spot anomalies in time-series data, said the crowdsourced traffic app Waze is using its real-time anomaly detection platform to improve app performance.

Waze said Tuesday (June 5) it would use the autonomous analytics platform to improve its trip-planning app by offering better real-time recommendations for routes and faster identification of glitches affecting users.

Anodot’s machine learning approach detects drops in app performance, correlating incident data with relevant metrics that allow operators like Waze to spot and fix anomalies as users seek the traffic path of least resistance.

In applications like the Waze route planner, those metrics can be affected by seasonal factors as well as time of day or week. Since most drivers use the app to plan their rush-hour commute, the partners stressed the importance of using machine learning to spot anomalies in challenging time-series data.

Anodot, Sunnyvale, Calif., said Waze would use its platform specifically to monitor key performance indicators such as how many drivers are using their app at a given time and usage trends in different markets. “Anodot is helping us find anomalies and identify issues that drivers may not be aware of, but we see when looking at the big picture,” said Orna Amir, manager of Waze’s analytics group.

Along with spotting anomalies, the collaboration is designed to track the performance of different Waze apps used for example on Android-based devices and iPhones. Local versions are often customized to account for traffic restrictions such as high-occupancy vehicles lanes and toll roads.

Amir said Waze would use the time-series platform to detect user preferences in different regions or underperforming app features.

David Drai, Anodot’s CEO and co-founder, noted that traditional business intelligence tools can take weeks to ingest data needed to spot anomalies and gauge performance. The company’s machine-learning approach is designed to handle huge data sets with a variety of inputs in real time. The goal for Waze’s crowdsourced GPS, map and traffic navigation platform is to boost app performance and improve route suggestions.

Anodot was founded in 2014 to address the unmet need for fast and accurate time-series analysis. The company initially worked with customers like the ride-sharing company Gett to help them make sense of data such as device usage and driver activity. The startup’s upgraded approach replaces manual scanning of anomaly reports with machine learning techniques designed to provide a real-time picture of how apps like Waze are performing. The approach allows users to address glitches as they occur rather than after rush hour ends.

Amir of Waze said beta testing of the Anodot platform helped it sort out real anomalies from false alarms. That capability “helps us to save time and stay focused” on actual app performance anomalies, Amir said.

Editor’s note: This article has been updated to correct the spelling of Anodot.

Recent items:

Inside Anodot’s Anomaly Detection System for Time-Series Data

Hidden Anomalies No Match for LivePerson’s Machine Learning Engine