Building an Agile Business Intelligence Platform at Zhongyuan Bank: Architecture, Practices, and Future Outlook

Platform Construction Goals – Zhongyuan Bank aims to satisfy diverse user groups—management, business staff, and technical analysts—by providing reliable, real‑time data querying, analysis, and exploration to support decision‑making.

BI Platform Development Timeline – From 2019 to 2022 the bank transitioned from third‑party reporting tools to a self‑built agile BI ecosystem, introducing online data extraction, low‑code smart table assembly, proprietary reporting systems, data dashboards, and a unified metric repository.

Agile BI Architecture – The platform integrates data sources (relational databases, data lakes, APIs, file imports) and delivers eight core capabilities: fixed‑report development, low‑code data processing, SQL analysis, visual analytics, data distribution, metric management, personalized portals, and data download.

Full‑Process Data Analysis – The workflow is divided into data preparation, low‑code or SQL‑based processing, analysis (fixed reports or self‑service dashboards), and data applications (dashboards, portals, screens).

Low‑Code Data Processing – Ten common operations (field setting, aggregation, filtering, derived fields, sorting, pivot, find‑replace, row‑to‑column, merges) are combined via a pipeline model to generate SQL automatically for business users.

Visualization Capabilities – The platform supports large‑screen dashboards and interactive visual boards using components such as ECharts and AntV, enabling rapid chart creation, composition, and sharing across devices.

Real‑Time BI Solution – Streaming data from Kafka is processed by Flink and persisted in StarRocks, providing second‑level analytics for use cases like management cockpits, marketing effectiveness, and risk alerts.

Intelligent Q&A Bot – User queries are matched against a knowledge base; if unmatched, similarity models and NL2SQL generate appropriate SQL to retrieve answers, demonstrating early AI‑driven assistance.

Future Outlook – Ongoing work focuses on enhancing platform capabilities, expanding user adoption, smoothing migration from legacy BI, and cultivating data‑analysis mindset. Future directions include AI‑powered insights, automated visualization, and opening APIs/SDKs for broader integration.