How Kuaishou Scales Intelligent BI: Insights from Its Data Platform

Introduction Kuaishou's Data Platform team aims to improve data decision‑making efficiency by building advanced computing engines and high‑performance data services, offering comprehensive analytics support for the business. As a top domestic data platform, Kuaishou's BI toolchain covers everything from basic data ingestion to advanced applications, providing self‑service analytics. The team continuously integrates multiple systems and AI technologies to explore intelligent scenarios, driving the construction and adoption of intelligent BI products. Facing diverse user needs, data quality issues, and technical challenges, Kuaishou proposes a controllable, trustworthy, and feasible solution framework that leverages AI to continuously optimize intelligent analysis capabilities, achieving low‑cost, large‑scale intelligent deployment.

Key Sections

Background Introduction

Challenges and Solution Approach

Solution Details

Application Practice

Future Outlook

Q&A Session

1. Background Introduction

Kuaishou Data Platform Department focuses on enhancing data‑driven decision efficiency (including analysis and experiment decisions) through advanced compute engines, high‑quality data warehouses, high‑performance data services, and a suite of data solutions, positioning itself among the top domestic big‑data platforms.

2. Kuaishou Data Platform Responsibilities

The department improves decision efficiency by providing advanced compute engines, high‑quality data warehouses, high‑performance services, and a range of data solutions that support both analysis and experiment decisions, helping the business grow.

3. Kuaishou Big Data Analytics Mission

The mission is to build an industry‑leading, one‑stop data analysis toolchain that offers comprehensive scenario solutions and boosts data‑analysis decision efficiency.

4. Architecture Overview

The platform consists of three layers:

Business Layer : Covers DA products for individual users, platform products for operations, and online services for main‑site and B‑side business.

Product Layer : Provides two types of analysis products—general BI (KwaiBI) and specialized analysis products for domains such as the main site and e‑commerce.

Service Layer : Supports two platforms—Gaia standard metric middle‑platform and API platform—offering data set query services, KV point‑lookup APIs, and SQL table query services.

5. BI Platform Introduction

Before diving into Kuaishou BI, the article defines Business Intelligence (BI) as tools or systems that transform complex data into actionable information for business decision‑making and planning.

6. Standard BI Process Flow

Data Ingestion: Connectors import data from various sources into the BI platform using unified table definitions.

Relationship Modeling: Define user‑friendly metrics and dimensions (e.g., GDP, province, date, city) based on the ingested tables.

Data Application: Perform analytical calculations (e.g., city‑level GDP distribution and anomaly trends) using the data set query service.

The article concludes with a discussion of future directions and a Q&A session.