Guide Hub

AI for Climate & Sustainability Masterclass

Climate work needs engineers as much as scientists. This masterclass is my working map of where software and machine learning actually help: risk models, early warning, satellite data, low-cost sensors, carbon accounting, and the energy cost of AI itself. It draws on Vayu, my award-winning air quality project, and it's honest about where the technology falls short.

24 lessonsStructured path
Your progress0/24 lessons

0% completed on this device.

Who this guide is for

  • Software engineers who want to work on climate and need the domain vocabulary.
  • Data and ML practitioners starting with satellite, weather, or sensor data.
  • Policy and sustainability people who want to understand what AI can and can't do for adaptation.

1. What Is Digital Climate Adaptation?

Mitigation gets the headlines. Adaptation is how people live with the climate we already have. What digital adaptation means, where data and AI help, and where they don't.

What Is Digital Climate Adaptation? →

2. Climate Risk Modelling Explained for Software Engineers

Hazard, exposure, vulnerability, scenarios. The vocabulary and architecture of climate risk models, explained for engineers moving into climate tech.

Climate Risk Modelling Explained for Software Engineers →

3. How AI Flood Early Warning Systems Work

From rain gauge to phone alert: how flood early warning systems work, where machine learning improves them, and where it still falls short.

How AI Flood Early Warning Systems Work →

4. Landslide Early Warning with ML

How rainfall thresholds, susceptibility maps, and machine learning combine into landslide early warning, using Nepal's monsoon season as a case study for mountain regions everywhere.

Landslide Early Warning with ML: A Nepal Case Study →

5. Impact-Based Forecasting

Telling people what the weather will do to them, not just what the weather will be. How impact-based forecasting and anticipatory action work, and what that means for engineers.

Impact-Based Forecasting: From Weather Data to Action →

6. Remote Sensing with Machine Learning

A developer-friendly introduction to machine learning on satellite imagery: bands, indices, labels, a first land-cover classifier in Python, and the traps that make results look better than they are.

Remote Sensing with Machine Learning: A Beginner's Guide →

7. Sentinel-2 vs Landsat for Environmental ML Projects

Resolution, revisit time, history, bands, and access: how to pick between Sentinel-2 and Landsat, and when to use both.

Sentinel-2 vs Landsat for Environmental ML Projects →

8. Best Python Libraries for Remote Sensing in 2027

The Python stack for satellite and geospatial work: reading rasters, handling time series, searching catalogues, vector data, ML, and visualisation, and how they fit together.

Best Python Libraries for Remote Sensing in 2027 →

9. Open Climate Datasets for ML Projects in 2027

A curated list of free climate and environmental datasets for ML projects: reanalysis, satellite imagery, land cover, forests, floods, air quality, and emissions, with what each is good for.

Open Climate Datasets for ML Projects in 2027 →

10. Low-Cost IoT Air Quality Monitoring

What I learned building Vayu, an IoT and remote-sensing air quality project that won a national award and went to APICTA: sensors, networks, data quality, and what I'd change.

Low-Cost IoT Air Quality Monitoring: Lessons From Vayu →

11. Calibrating Low-Cost PM2.5 Sensors Against Reference Data

Why low-cost PM2.5 sensors over-read in humid air, how to co-locate with a reference monitor, and how to fit and validate a correction model you can trust.

Calibrating Low-Cost PM2.5 Sensors Against Reference Data →

12. Forecasting Air Quality with Machine Learning

How to forecast PM2.5 a day ahead with machine learning: features that matter, a gradient boosting baseline, honest evaluation, and why weather data does most of the work.

Forecasting Air Quality with Machine Learning →

13. Building a Climate Data Dashboard with Next.js and MapLibre

A frontend engineer's guide to climate dashboards: MapLibre in Next.js, raster and vector layers, time sliders, performance with large datasets, and designing for people who make decisions.

Building a Climate Data Dashboard with Next.js and MapLibre →

14. How Carbon Accounting Software Works Under the Hood

Carbon accounting is activity data times emission factors, plus a lot of plumbing. The data model, the calculation engine, and the hard parts, explained for engineers.

How Carbon Accounting Software Works Under the Hood →

15. Scope 3 Emissions Data

Scope 3 is usually the biggest share of a company's footprint and the hardest to measure. Why the data is so messy, the three calculation methods, and how software teams improve it over time.

Scope 3 Emissions Data: Why It's So Hard to Collect →

16. Best Carbon Accounting Software in 2027

How to evaluate carbon accounting platforms like an engineer: methodology, data integrations, Scope 3 support, audit trails, and reporting frameworks, plus the vendors buyers commonly compare.

Best Carbon Accounting Software in 2027: How to Compare Platforms →

17. How to Calculate a Carbon Footprint (Methods Compared)

Every carbon footprint calculator multiplies what you do by an emission factor. How the maths works, why calculators disagree, and how to calculate your own footprint transparently.

How to Calculate a Carbon Footprint (Methods Compared) →

18. How Much Energy Does AI Use? The 2027 Trade-offs

AI's energy use is real and growing. So are its climate benefits in some places. How to think about the trade-off honestly, and what developers can do about their own footprint.

How Much Energy Does AI Use? The 2027 Trade-offs →

19. Responsible AI in Climate Policy

When governments use AI to target climate support, issue warnings, or plan adaptation, mistakes land on real people. The risks that matter and the guardrails I'd put in place.

Responsible AI in Climate Policy: Risks and Guardrails →

20. Best Climate Tech Tools and APIs for Developers in 2027

The open-source libraries, data platforms, and APIs I'd reach for when building climate software: earth observation, weather, emissions, energy, and geospatial tooling.

Best Climate Tech Tools and APIs for Developers in 2027 →

21. European Electricity Prices per kWh in 2027

How European electricity bills are built (unit rates, standing charges, taxes, dynamic tariffs), where to find current prices, and how to calculate what an appliance actually costs.

European Electricity Prices per kWh in 2027 →

22. UK Energy Price Cap 2027

What the Ofgem price cap actually caps, why the 'typical household' figure isn't your bill, and how to calculate your own cost from unit rates and standing charges.

UK Energy Price Cap 2027: Calculate Your Bill →

23. Heat Pump Running Costs

How to estimate what a heat pump costs to run: heat demand, COP and SCOP, electricity vs gas prices, and the break-even ratio that decides whether it saves money.

Heat Pump Running Costs: The kWh Maths →

24. EV Charging Cost at Home vs Public Chargers

How to calculate what it costs to charge an EV at home and on public chargers, per charge and per mile or kilometre, including charging losses and time-of-use tariffs.

EV Charging Cost at Home vs Public Chargers →

Frequently asked questions

Do I need a climate science degree to work on climate AI?

No. Most climate tech teams need engineers who can build reliable data pipelines, models, and interfaces. Domain knowledge matters, and this guide covers the vocabulary you need to start.

What is the difference between climate mitigation and adaptation?

Mitigation reduces emissions to limit future warming. Adaptation helps people and systems cope with climate impacts that are already happening, for example through early warning systems and risk mapping.

Where can I get free climate and satellite data?

Sentinel and Landsat imagery, ERA5 reanalysis, and many forest, flood, and air quality datasets are free. The guide lists them with what each is good for.

Is AI bad for the climate?

AI uses significant energy, especially in data centres, but it can also cut emissions and harm elsewhere, for example in grid optimisation and early warning. Whether it's worth it depends on what it's used for and how it's powered.

Contact

Get in Touch

Want to call me? book a call and I'll respond in the available time slot. I will ignore all soliciting.