Beacon: Revolutionizing Media Analytics With Near Real-Time Declarative Pipelines
Overview
| Experience | In Person |
|---|---|
| Track | Data Engineering & Streaming |
| Industry | Enterprise Technology, Communications - Media & Entertainment |
| Technologies | Lakeflow, Unity Catalog, Lakebase |
| Skill Level | Intermediate |
CondéNast operates in a digital first media landscape where editorial and business decisions are defined in minutes. To meet this need, we built Beacon, a near real time media analytics platform that transformed a fragmented, high latency analytics stack into a unified architecture powered by the Databricks Lakehouse and Lakeflow Spark Declarative Pipelines.Having InfluxDB and Qlik solution, introduced a 15 minutes data lag and inconsistent KPIs. It was re-architected by streaming data directly from web collectors into Databricks, unifying streaming and batch processing through declarative pipelines. A Redis-based caching layer enables sub second access while preserving the Lakehouse as the system of record.This approach reduced end-to-end latency to under 10 seconds, enabling editorial, revenue and growth teams to monitor article performance, live events and monetisation signals in near real time through a single analytics experience. This will be enabled with Genie LLM in future roadmap.
Session Speakers
Jitendra Sharma
/Senior Data Engineering Manager
Conde Nast Publications
arun karthik
/Director
Conde Nast