Application of Multi‑Dimensional Analysis in NetEase Cloud Music's Social Innovation Business
The talk explains how NetEase Cloud Music leverages multi‑dimensional analysis across its social innovation apps, detailing business background, use‑case scenarios, technology choices, data‑platform architecture, and future plans to enable self‑service analytics and improve data efficiency.
01 Business Background Introduction
NetEase Cloud Music's innovation business consists of a matrix of social apps—such as the stranger‑social app "XinYu", voice‑interaction app "ShengBo", and overseas app "HeatUp"—each targeting different user groups. All share commercial goals, rapid iteration, and the need for efficient data support, prompting the development of a multi‑dimensional analysis solution.
02 Multi‑Dimensional Analysis Application Scenarios
The evolution of data support is described in four stages: (1) initial data‑warehouse construction using Hive and Spark; (2) manual data extraction and Excel analysis; (3) a report system with SQL‑based self‑service analysis; and (4) the current stage where a dedicated multi‑dimensional analysis platform (easyFetch) integrates Impala, Kylin, ClickHouse, and Greenplum to provide transparent, unified access to diverse storage back‑ends.
Technical selections progressed from Hadoop‑based offline warehouses to real‑time pipelines built on Kafka and Flink, forming a Lambda architecture and exploring batch‑stream convergence.
03 Significance of Self‑Service Multi‑Dimensional Analysis
Self‑service analysis enhances data value by supporting decision‑making, marketing activities, and post‑event effectiveness evaluation. It enables business users to explore data freely, generate ad‑hoc reports, and convert analysis results into SQL for deeper investigation or into user segments for targeted push or SMS campaigns.
04 Data Foundation for Multi‑Dimensional Analysis
The platform relies on rich thematic models built on a "person‑product‑place" abstraction, allowing modular dimension tables (person, product, place) and aggregated tables at various granularities. Data assets are managed like products, with white‑papers documenting models, training, feedback loops, and continuous improvement.
05 Future Vision
While current model construction still depends on data‑warehouse engineers, the roadmap aims to empower users to build their own multi‑dimensional models. This will introduce challenges in model lifecycle management, diversity control, and value assessment, which will be addressed through supporting tooling and automated evaluation mechanisms.
Overall, the presentation highlights how a robust multi‑dimensional analysis platform can accelerate innovation, reduce reliance on data engineers, and create a virtuous data‑driven loop for NetEase Cloud Music's social products.
DataFunSummit
Official account of the DataFun community, dedicated to sharing big data and AI industry summit news and speaker talks, with regular downloadable resource packs.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.