Learning path
GeoAI: from zero to your first map
Learn the basics, work with free imagery and finish with a land cover map you tested honestly.
- Length
- 7 articles · about 63 min
- Level
- Beginner to intermediate
- Before you start
- None to start. Basic Python helps from the NDVI tutorial onwards.
Best for: Analysts, planners and students new to GIS or satellite data
By the end
What you will be able to do
- Explain rasters, vectors and coordinate systems in plain language
- Choose suitable free satellite imagery for a question
- Compute and cloud-mask a vegetation index from Sentinel-2
- Train and honestly test a land cover classifier
- Share results on an interactive web map
- Scope a GeoAI pilot with a clear decision and success measure
The path
Read in this order
Concept · Beginner · 6 min read
GIS basics: rasters, vectors and coordinate systems
The three ideas every beginner needs before opening any GIS software, explained without jargon.
You will learn
- Vector data stores shapes (points, lines, polygons). Raster data stores a grid of values, such as pixels.
- A coordinate reference system (CRS) tells software how numbers map to places on Earth.
Start here →
Concept · Beginner · 8 min read
GeoAI in plain language
What GeoAI is, what it can and cannot do, and how to start a first project without getting lost in jargon.
You will learn
- GeoAI applies machine learning to data that has a location: imagery, maps, GPS traces and sensor readings.
- It is strongest at repetitive mapping tasks across large areas, where people cannot cover every location by hand.
Read →
Concept · Beginner · 7 min read
Resolution in satellite imagery: GSD, revisit time and bands
The numbers that decide whether an image can answer your question, and how to choose between free and commercial data.
You will learn
- Spatial resolution (GSD) is the size of one pixel on the ground, and it sets the smallest thing you can see.
- Revisit time is how often a satellite images the same place, and cloud cover reduces it in practice.
Read →
Tutorial · Beginner · 11 min read
Compute NDVI from Sentinel-2 imagery with Python
Turn the red and near-infrared bands into a cloud-masked vegetation health map, step by step.
You will learn
- NDVI = (NIR − Red) / (NIR + Red), with values from −1 to 1.
- Sentinel-2 band B08 is near-infrared and B04 is red, both at 10 m resolution.
Read →
Tutorial · Intermediate · 13 min read
Classify land cover with a random forest in Python
Train a machine learning model on labelled pixels, test it honestly and turn a satellite image into a land cover map.
You will learn
- Each pixel becomes one row of data, and its band values become the features.
- A random forest is a strong, easy first model for land cover classification.
Read →
Tutorial · Beginner · 12 min read
Show your data on a web map with MapLibre
Add an interactive map, a GeoJSON layer, popups and data-driven colours to any web page.
You will learn
- MapLibre GL JS is a free, open-source library for fast interactive vector maps.
- GeoJSON is a simple format for points, lines and polygons, and it uses [longitude, latitude] order.
Read →
Guide · Intermediate · 6 min read
How to scope a GeoAI project
A seven-step framework that turns a vague idea into a testable pilot, with the questions to settle before any model is built.
You will learn
- Define the question as a decision, then the area, season and update frequency.
- Check that imagery and labels exist before you promise any level of accuracy.
Read →
When you finish
Where to go next
Next path
Back to basics: the maths and science behind GeoAI, 3D and VR
AI can draft the code, the model and the scene. People are still needed to develop, fine-tune and put them in context, and that takes the fundamentals.
10 articles · 66 min
Next path
RAG: from idea to a trusted assistant
Understand how it works, split documents well, build a pipeline and prove it answers correctly.
5 articles · 37 min
Start a project
Want this applied to your own project?
Learning is the first step. If you would rather have a working system, book a discovery call and we will scope it with you.