Tencent PCG Data Governance System Overview

The presentation introduces Tencent PCG's data governance system, outlining the current problems of massive, complex data and fragmented technical stacks, which lead to high‑cost, low‑efficiency "movement‑style" governance.

To address these issues, the solution is divided into four chapters: (1) an overview of data governance concepts and approaches; (2) PCG meta‑warehouse construction focusing on feature extraction and governance engine foundations; (3) an open, sustainable asset‑scoring system; and (4) a one‑stop governance workbench.

For managers, the platform provides a macro view of resource utilization, identifies governance opportunities, assists in strategy formulation, and tracks execution effectiveness, turning ad‑hoc governance into a continuous process.

For data producers, the platform lowers governance thresholds by offering one‑click operations such as lifecycle adjustment and backup automation, thereby reducing storage costs and simplifying routine tasks.

The meta‑warehouse layer integrates full‑link data cleaning, lineage construction, and feature mining, supporting multiple script types (SuperSQL, Thive, Hive, Spark) through a unified parsing engine that abstracts syntax differences.

Asset scoring evaluates assets across five dimensions—norms, cost, security, quality, and application—assigning scores that reflect health and guiding KPI definition.

The governance engine transforms extracted features into actionable governance items via a rule engine, applies scoring rules, and delivers personalized governance plans to managers and data owners.

The one‑stop governance workbench enables managers to view asset dashboards, define and push governance plans, allow data owners to execute tasks (e.g., table deletion), and review governance outcomes, including cost savings and asset‑score improvements.