Data Value System and Cockpit Construction: A Case Study from CITIC Bank

Introduction

The deep integration of finance and technology has made digital transformation a strategic imperative for banks, prompting the development of data value systems and cockpit applications to drive business insight and operational efficiency.

01 Data System Business Goals

The bank defines its data system goals around four questions: What (quickly answer business status and future trends), Why (analyze root causes), How (leverage AI and data mining for rapid experimentation), and Do (execute, monitor, and close the loop). Agile processes and knowledge management are emphasized to support this closed‑loop.

02 Overall Planning of the Data System

The architecture consists of one foundational data layer (data lake, data warehouse, real‑time processing platform), one governance system covering data standards, metadata, lifecycle and quality, two toolsets (BI tools for visualization and reporting, AI tools for model development and deployment), and six major application domains (digital management, intelligent marketing, intelligent risk control, smart asset‑liability, smart finance, and regulatory technology).

03 Digital Management System

This subsystem provides a core decision‑support platform for the whole bank, addressing unclear indicator logic, non‑intuitive traditional reports, inconvenient data access, and untimely data publishing by delivering visual dashboards and mobile‑friendly cockpits.

04 Methodology and Case Demonstration

The methodology includes background analysis, definition of the data value system, analytical thinking, implementation steps, and asset delivery. It stresses cross‑team collaboration (business, analysis, data engineering, real‑time development, front‑end) and outlines 15 standard implementation stages from requirement analysis to production and operation.

Key case studies include the development of 11 mobile cockpits in 2020 for headquarters, branches, and sub‑branches, as well as a retail cockpit covering performance, analysis, real‑time monitoring, and process control. The retail cockpit’s indicator tree illustrates how metrics such as AUM growth, effective calls, and conversion rates are broken down and monitored.

05 Cockpit Deployment and Usage

As of the latest data, the retail cockpit has over 4,000 active users per month with approximately 160,000 page views. Additional branch‑level cockpits are being promoted to support daily management and performance analysis.

Overall, the presentation demonstrates how a structured data value system and cockpit implementation enable banks to achieve data‑driven decision making, improve operational agility, and continuously iterate on business metrics.