The GeoAgent: Natural Language EO Queries

Turning a plain-English request into a full analysis pipeline — offline or LLM-assisted

The GeoAgent: Natural Language EO Queries
Intermediate Geospatial AI 2 lessons Samuel Appiah Kubi

About this course

Every pipeline in this curriculum has the same underlying shape: pick a sensor, pick a task, run an ordered sequence of tool calls. GeoAgent automates exactly that decision, so a disaster-response officer or agricultural extension worker can type a plain English sentence instead of writing Python — and it does this with two different planners, one that needs no external API at all.

What you'll learn

  • Run flood detection, crop mapping, and change detection with a single English sentence
  • Understand GeoAgent's 3-stage decision hierarchy: sensor, then task, then tool sequence
  • Explain the trade-off between the offline heuristic planner and the LLM-backed planner
  • Run GeoAgent with zero external API key using heuristic mode
  • Extend GeoAgent's decision matrix with custom tools for your own organisation

Requirements

  • Satellite Data Acquisition with pygeofetch
  • Basic familiarity with the pygeovision client

Course content

Your First Natural-Language Analysis Preview 24 min
The 3-Stage Decision Hierarchy, and Why Two Planners Exist 20 min
Free
Login to Enroll
  • LevelIntermediate
  • Lessons2
  • CertificateYes
  • AccessLifetime
Samuel Appiah Kubi Instructor