Quick Sign-Up
Just so we know who's exploring, and so you can pick up where you left off. Takes 10 seconds, no password needed.
We only use this to track course progress and issue certificates. No spam, ever.
Just so we know who's exploring, and so you can pick up where you left off. Takes 10 seconds, no password needed.
We only use this to track course progress and issue certificates. No spam, ever.
Air pollution isn't just uncomfortable, it's a real health risk, and some people feel it more than others. In four short stops, you'll learn what's actually in the air, check live air quality anywhere in the world, see how satellites help us track it, and ask a real AI assistant about it yourself.
Bad air isn't just unpleasant, it can genuinely affect health. And it doesn't affect everyone the same way.
Here's the short version of why: the air you breathe carries tiny particles called PM2.5, so small they slip past your body's usual defenses and travel deep into your lungs, and from there into your bloodstream. Breathe in enough of them, often enough, and your body pays for it, your lungs and heart have to work harder, and if you're pregnant, it can affect the pregnancy too. None of that is visible in the moment. That's exactly why it needs data and science to see clearly, not just how the air looks or smells.
Air pollution is linked to roughly 7 million early deaths a year worldwide, and household air pollution from things like cooking smoke accounts for more than 4 million of those on its own, according to the World Health Organization. This isn't a rare or distant problem, it's one of the biggest environmental health risks on the planet.
This isn't hypothetical, it's happening in neighborhoods across Africa right now. Place Alert Labs researchers study exactly this, in real cities, to figure out where to act first.
The smoke from indoor cooking fires can sometimes make the air inside a home worse than the air outside. It's not just about traffic and factories.
Nobody can put a sensor on every street. So GeoAI researchers train models to guess where air is likely worst, combining things like traffic density, cooking fuel type, and satellite haze readings, before a single sensor is ever installed there. It's less "detective with a magnifying glass" and more "pattern-spotter that checks every neighborhood at once, and never gets tired."
Which situation probably means breathing in more polluted air overall?
Optional, no right answer, nobody's grading this:
Five simple terms are all you need to talk about air quality like a scientist. Once you know these, every air-quality app, news report, and research paper starts making a lot more sense.
The single most important measurement to track is your PM2.5 level, short for particulate matter 2.5 micrometers or smaller. It's what most air-quality guidelines, indexes, and the rest of this course are built around, so if you only remember one term, make it this one.
Here's roughly where a few real-world PM2.5 readings would land on the scale. The line marks the WHO guideline.
Anything past the line is considered a health concern by WHO standards.
Check your own exact spot, live. Use your device's location, or type any city or town.
Live data: Open-Meteo Air Quality API, ground and satellite-blended estimates, updated hourly.
A satellite is just an object orbiting Earth, carrying instruments, not a spy camera. Some carry sensors that measure how much sunlight makes it through the atmosphere.
Most places, especially across Africa, don't have ground air-quality sensors on every street. Satellites give us coverage everywhere, even where there's no equipment on the ground at all.
The catch: satellites don't measure PM2.5 directly, they measure AOD, and turning AOD into an estimated PM2.5 number takes calibration against the few real ground sensors that do exist. That's actual GeoAI work: teaching a model the relationship between what the satellite sees and what a ground sensor reads, then using that model to fill in every gap in between.
More particles → less light reaches the satellite → the satellite reads that as a higher AOD, which gets turned into an estimated PM2.5 value.
Move the slider until the estimated AOD reads about 0.50.
Give it a try!
Ground sensors are accurate, but they're expensive to buy, need power and upkeep, and only tell you about the one spot they sit in. Covering a whole country with them isn't realistic. A satellite already up there, passing overhead, gives rough estimates everywhere at once, even in places nobody has ever installed a sensor.
A single AOD reading isn't very useful by itself. Haze, dust, and humidity all block sunlight in slightly different ways. This is where GeoAI comes in: models are trained on the rare spots where a satellite reading and a real ground sensor exist side by side, learn the relationship between them, then apply that same pattern to every satellite pass, even over places that have never had a sensor.
Now build your own comparison. Add up to 5 real places, your home, your school, a city you have only ever heard about, anywhere in the world, and we will pull live PM2.5 air quality for each one right now. We have already added two real reference points so you have something to measure against: Lahore, Pakistan, one of the most polluted cities on Earth, and Reykjavík, Iceland, one of the cleanest.
Ask a question in plain English below. A real AI assistant reads whatever you have added and answers you, the same way GeoAI researchers question their own datasets.
Add at least one place above to see live data and unlock the map.
Trend column: each line is 5 days recorded + 4 days forecast for that place, left to right. The gap in the dashes marks today. See Stop 2's live check for the same chart with numbers.
These steps complete themselves as you actually do them, no clicking required: