Agentic Coding for Geospatial

This course covers agentic coding workflows specializing in geospatial data science. We cover best practices for using coding agents to rapidly solve complex problems - grounded in established science and human-in-the-loop validation. We cover hands-on examples using Claude Code – how to use Skills and MCP servers to augment domain-specific capabilities.

Ways to take this course

Every course is free to study on your own. The live classes add certification, support and a cohort to work through it with.

Self study

Free, at your own pace

Pre-recorded videos
Full course material

Free

Start learning

Live classes

Cohort-based, with certification

Live instructor-led Zoom sessions
Certificate of completion
Lifetime technical support
Mentoring

Next class: Nov 16–24, 2026

US$175 · ₹12500

Notify me when it opens

Learning outcomes

Directing AI coding agents inside a real dev environment to build, debug, and iterate on geospatial code

Configuring and extending agent capabilities using skills and MCP servers for domain-specific workflows

Building and executing cloud-based notebooks for compute-heavy tasks like deep learning inference

Course outline

Module 1

Agentic Coding Basics

Introduction and Course Overview
Introduction to Agentic Coding
Geospatial Data Analysis (Mapping Crime Hotspots)
Exploratory Data Analysis (EDA)
Data Cleaning and Spatial Analysis

Module 2

Building Complex Workflows

Using Web APIs (Network Analysis and Route Optimization)
Using Claude Code in VS Code
Building Notebooks
Interactive Data Analysis
Validation and Export

Module 3

Building a Personal Knowledge Website

Building a Personal Knowledge Website
Setting up your Personal Knowledge Base
Populating Notes with Obsidian
Publishing Your Notes Website

Assignment

Calculate Livability Score

Module 4

Skills and MCP Servers

Agent Skills
Using Skills
Creating a Skill
Using MCP Servers
Using QGIS MCP Server (Multicriteria Overlay)

Module 5

GeoAI

Building Notebooks in the Cloud
Using Deep Learning Models

What participants say

Read moreShow less

Agentic Coding for Geospatial gave me a working method. During the course I used an AI coding agent to build a complete building-detection pipeline end to end in a Colab notebook: tiling a very high-resolution drone orthomosaic (3cm), reprojecting each tile to its local UTM zone, resampling to match a pre-trained GeoAI model’s training resolution, running inference, and saving the results as GeoJSON. When the pipeline produced zero detections partway through, the agent diagnosed the actual cause, tiles that had shrunk below the model’s working size after resampling, rather than papering over it, and fixed it. That is the kind of practical, tool-grounded workflow I was looking for. I recommend the course to anyone doing applied GIS work who wants to move faster without losing rigor.