Training

Hands-on geospatial data science and GeoAI, taught in a room. From the fundamentals to production workflows, and from open public workshops to programmes built entirely around your team's field and data. One rule holds across all of it: you leave with code that runs on your machine, not a notebook you watched someone else run.

Two formats

In-house training on a standard curriculum

One or two days · Your team · Your office or remote

The same modules, delivered to one organisation's team so everyone starts from the same place and finishes with the same toolkit. Pick the modules that match where your team is:

• Spatial data fundamentals — GeoPandas, coordinate systems, spatial joins, and the mistakes that silently return nothing.

• Urban analytics — OpenStreetMap and Overture, H3 hexagons, accessibility and catchments, comparing neighbourhoods honestly.

• Satellite data science — Sentinel and Landsat time series, cloud masking, change detection, building a clean signal from noisy scenes.

• GeoAI — machine learning on spatial features, validation against ground truth, knowing when the model is right for the wrong reason.

• Network science for spatial systems — mobility, connectivity, and flows as networks.

• Maps that persuade — data visualisation and cartography for decision-makers, not for other analysts.

Tailored corporate programmes

Scoped to your field, your data, your questions

For teams that need more than a generic curriculum. We start from what your organisation actually decides with location data — sites, portfolios, risk, markets — and build the day around it: your datasets where they can be shared, your domain's open data where they can't, and exercises that end in something your team can use the following week. Delivered to real estate, energy and infrastructure, insurance, public-sector and economics teams.

Pricing

Company trainings start from a half-day online at €2,800 + VAT. The final price depends on whether it runs on-site or online, the exact topics, the size of the group, and any custom work on your data or domain before the day. Multi-day programmes and follow-up sessions are priced on scope. For a quote, reach out with the two things below.

How a day runs

For years everything I've taught went through a screen — courses, books, video. There's a version of teaching that doesn't survive compression: standing behind your shoulder while your spatial join fails, fixing it together, and moving on. That's the version these days are built on.

• Your own laptop, your own environment. Set up in about 30 minutes of pre-work the week before, so the day starts with data, not with installs.

• Real open data, real models. Nothing synthetic. What you build runs after you leave.

• A repo you keep. Everything from the day, structured to extend.

• The debugging done with you. Environments, CRS mismatches, joins that return zero rows — fixed in the room, not in a comment thread afterwards.

• Small rooms. Twenty people at most, one instructor, time for the specific question about your actual problem.

Who it's for

Working data analysts, data scientists, engineers and GIS professionals who want to ship a spatial workflow, not watch one. You need Python at a daily-driver level — functions, pandas, comfortable in a notebook. You don't need prior GIS; the fundamentals block exists so nobody does.

Please self-select honestly. A room split between beginners and practitioners fails for both — which is exactly why the tailored format exists for mixed teams.

Who's teaching

Milan Janosov, PhD — network and geospatial data scientist, Budapest. A decade of teaching this material three ways: full-day in-house workshops for corporate teams; TEDx and conference keynotes on cities, networks and what spatial data can actually tell us; and as visiting faculty at Central European University and MOME Open. Alongside that, seven online courses, three books, and a research background through the Barabási Lab and Bell Labs. Over 60,000 people have taken the courses and books; more than 130,000 follow the work online.

Bring a team

For an in-house day or a tailored programme, send me two things:

1. Who's in the room — team size, roles, and where they are with Python and GIS.

2. What they should be able to do the Monday after — the workflow, the dataset, the decision.

Scope a training →