From Map to Agent

Overview

From 2020 to the present, I've led the full UX evolution of Tudou Data's geospatial platform — a national-level system serving government clients including the Shaanxi Provincial Natural Resources Bureau. The platform evolved from a 3D spatiotemporal data visualization tool into a comprehensive AI large-model Agent OS, recognized as a 2025 Digital Intelligence Leading Achievement.

The Problem

Government-scale geospatial data is inherently complex — vast datasets, multi-dimensional visualization requirements, strict regulatory standards, and users ranging from field surveyors to provincial bureau directors. The challenge was not just making data visible, but making it actionable at every level of the organization.

As AI capabilities matured, the platform needed to evolve from a passive visualization tool into an intelligent agent system — without losing the precision and reliability that government clients require.

2025 — AI Agent OS

The 弗雷AI工厂 (Frey AI Factory) platform represents the culmination of this evolution: a multi-agent OS where government users interact with geospatial intelligence through natural language. Agents handle tasks ranging from map diagnostics and document validation to hydrological monitoring and land change detection.

Key Outcomes

• ◆Led national-level project 'Qinling Natural Resources Survey Monitoring Pilot' — delivered to Shaanxi Provincial Natural Resources Bureau with high recognition
• ◆Designed and iterated the 3D spatiotemporal database UX across 6 years of platform evolution
• ◆Architected the UX for the AI large-model Agent OS platform, enabling government users to interact with geospatial intelligence through natural language
• ◆Recognized as a 2025 Digital Intelligence Leading Achievement