Case study
Real-Time Learning Analytics on Looker and GCP
How SquareShift replaced a nightly batch BI process with real-time Looker and GCP analytics for an EdTech learning platform.
Dashboards feel real-time until someone checks the web page against them. The fix isn't a faster nightly batch — it's one pipeline feeding both from the same data.
Looker and BI Practice Lead, SquareShift
A global EdTech company delivering an AI-powered learning experience platform to enterprises and government organizations at scale.
Its learners and L&D managers needed real-time insight the existing nightly-batch BI stack couldn't provide.
- Real-time insights — Looker replaced nightly batch reporting for learners and L&D managers.
- Embedded analytics — The same Looker reports now run directly inside the web experience.
- Unified golden data — Web page numbers and BI reports finally match.
How SquareShift delivered it.
The challenge
Learners and L&D managers needed to see activity and progress as it happened, but BI reporting ran on a nightly batch job — always a day behind. The data science team’s models had no real-time data to work with either, and the numbers shown on the web page never matched what the BI reports said, undermining trust in both.
What we delivered
SquareShift’s Looker-certified solution architects built a serverless blueprint architecture with the client, replacing the nightly batch job with real-time pipelines: Kafka streamed data from every source, and Spark handled enrichment. Looker then served near-real-time insights, embedded directly into the web experience rather than living in a separate reporting tool.
The payoff
Learners and L&D managers now see their activity and progress in real time, on the same golden data that feeds the web experience. Web numbers and BI numbers finally agree, closing a gap that had never been resolved before.
Where this work sits