How Tencent Powered China’s 7th Census with Big Data and Cloud Tech

On May 11, the results of China’s seventh national census were announced, showing a total population of 1,411.78 million , which continues to grow.

The census is directly related to every Chinese citizen, and its data collection required massive technical effort, including nationwide data gathering, storage, transmission, de‑identification, analysis, and verification.

First full‑scale electronic census, allowing investigators to report data in real time.

First use of self‑service online reporting via the internet.

First application of administrative big data for cross‑checking.

First use of cloud computing and cloud services for real‑time processing.

First information system managing over 7 million census workers online.

To enable nationwide self‑reporting, a WeChat mini‑program was built for citizens to scan a QR code and fill in their information on mobile devices. Another mini‑program served the 7 million volunteers, providing task assignment and data collection functions.

Tencent’s technical support included cloud databases capable of handling billions of user records, millions of terminals, and peak TPS in the millions . The TDSQL database achieved a maximum throughput of 2.5 million transactions per second , processed tables with over 2 billion rows , and sustained data transmission rates of up to 500 k rows per second (average 300 k rows/s) . Query per second (QPS) peaked at around 110 k on the first day of the census.

The system leveraged a dual‑engine design: an OLTP engine for high‑performance transactional processing and an OLAP engine for massive analytical workloads, supporting petabyte‑scale data analysis. High availability was ensured with a primary‑plus‑three‑replica architecture and cross‑region disaster‑recovery.

All development was performed using cloud‑native containerization (TKE) and micro‑service architecture, eliminating the need for manual server setup and allowing the system to support 100 000 concurrent submissions . Security measures such as micro‑service gateways, authentication, and permission controls protected service quality.

Compared with the 2010 paper‑based census, the electronic approach reduced the registration time to under 20 minutes , dramatically lowering post‑collection data entry effort.

Over 300 Tencent engineers from more than 40 departments participated for more than 200 days, contributing the technical backbone that made this massive national survey possible.