Case Study: The Globe and Mail delivers real-time article recommendations with ClearScale

The Big Data Infrastructure That Powers The Globe and Mail’s Article Recommendation Engine

The Globe and Mail, Canada’s #1 weekday and weekend newspaper, wanted to add a personalized article recommendation system to its new mobile app. The company had already prototyped several algorithms in Amazon, but it needed ClearScale to help architect a real-time production infrastructure in AWS within tight timelines.

ClearScale built the AWS infrastructure behind the recommendation engine using services including Kinesis, Elastic MapReduce, DynamoDB, S3, CloudFormation, OpsWorks, and Chef, along with a custom data transfer application. The system delivers next-article recommendations by the time a reader loads the current story, and The Globe said the custom solution gave it more potential than off-the-shelf recommenders while supporting future expansion to web and mobile web with no system changes beyond throughput and cluster size adjustments.

The Globe and Mail

Michael O’Neill

Director of Data Science