AI Case Study

Coyote improves the accurate reporting of speed limits by 9% using machine learning

Coyote provides real time information on traffic hazards, road conditions and speed limits. In order to maintain an accurate reporting of driving speed limits within their embedded maps, the company implemented Dataiku's machine learning solution. By leveraging predictive algorithms the company was able to detect speed limit anomalies which led to a 9% increase in speed limit reliability and the automation of its correction process.

Industry

Transportation

Freight And Logistics

Project Overview

"Armed with Dataiku DSS, Coyote used its Machine Learning capabilities to detect anomalies in their speed limit referential, within specific datasets. Coyote developed an algorithm to leverage vast amounts of IoT-derived data. It segments roads into sections and analyze patterns in each section. This enabled Coyote to build a predictive model that estimated the speed limit of the road section. The Machine Learning process facilitated the detection of speed limit anomalies and, consequently, enabled Coyote to estimate the global quality & reliability of the displayed speed limit. The entire process wouldn’t have been possible without Dataiku DSS’ collaborative functionalities. Thanks to the platform’s focus on teamwork & cooperation, which enabled employees with differing skill-sets to work together, Data Mining & Visualization are now widespread within the company and there is a growing awareness of Smart Data issues."

Reported Results

Coyote achieved significant improvement of speed limit referential.
"Using Dataiku Data Science Studio enabled Coyote to improve their core product’s accuracy & efficiency while establishing a data-driven spirit within the entire company."

Key accomplishments include:

* Increase of 9% for speed limit reliability on analyzed datasets
* Automation of the speed limit correction process
* Increased customer loyalty

Technology

"Coyote developed an algorithm to leverage vast amounts of IoT-derived data. It segments roads into sections and analyze patterns in each section. This enabled Coyote to build a predictive model that estimated the speed limit of the road section. The Machine Learning process facilitated the detection of speed limit anomalies and, consequently, enabled Coyote to estimate the global quality & reliability of the displayed speed limit."

Function

R And D

Product Development

Background

"Coyote is the European leader of real-time road information. Coyote uses IoT-based devices and mobile applications that enable their users to warn other drivers of tra ic hazards and conditions (e.g., traffic obstruction, accident, speed camera, etc.) that are detected while driving."

The company wished to improve speed limit reliability. "Coyote’s IoT devices and apps rely heavily on the accuracy of incoming data. Of particular interest are the driving speed limits within their embedded maps. Keeping them accurate and up-to-date is a big challenge for Coyote’s quality teams. In terms of data analysis, Coyote needed an automated algorithm-based solution that would correct speed limit data. Ideally, the solution would leverage the high volume of incoming data from their IoT devices (billions of rows with anonymized speed and position of their users) to turn them into actionable insights and predictions. By association, this also meant that Coyote needed to instill a data-driven approach within the company — decisions needed to be based on real-world data rather than standards analytics reports."

Benefits

Data

IoT-derived data, speed limits