# How Google Maps Works in Real Time How Google Maps finds your position from satellites, builds its street map and predicts traffic, so your route and arrival time update as you drive\. By AshokR Topic: Computing, AI & Electronics [Open the interactive Jig](https://ask.jig.com/j/yQ95OTbAVdSz) ## Transcript ### 1. A map that is alive Google Maps is more than a picture of streets\. It keeps blending where you are, how fast traffic is moving and what usually happens next\. Let's follow one trip through this little city, one layer at a time\. More than 2 billion people use Google Maps every month\. Drag the city to turn it\. ### 2. Where am I? Ask the sky Your phone listens to GPS satellites about 20,000 km above Earth\. Each one sends a timed signal, and by comparing signals from at least four, the phone works out where it is\. The satellites only send: they never learn where you are\. GPS stands for Global Positioning System\. The satellites here are drawn much closer than they really are\. ### 3. Filling the gaps Tall buildings and tunnels can block satellite signals\. Then the phone also uses nearby Wi\-Fi and cell towers, plus motion sensors that feel turns and speed\. Maps snaps the blue dot onto the most likely road, which is called map matching\. The pale blue circle shows how unsure the phone is about your exact spot\. ### 4. The map underneath The streets themselves come from many sources\. Street View cars photograph the roads, and computers read street names, house numbers and shop signs in the pictures\. More than 1,000 official data sources and millions of daily fixes from people keep it up to date\. Street View holds more than 280 billion photos from over 100 countries\. They are snapshots, not live video\. ### 5. Millions of tiny speedometers Phones with location turned on share anonymous speeds and positions\. Many readings are mixed together, so no single person is shown\. When lots of phones crawl along a road it turns orange or red; free\-flowing roads stay green\. Google started this crowd\-sourced traffic on Android phones in 2009\. You can switch location sharing off in your settings\. ### 6. Guessing the next half hour Live traffic shows right now, but your trip happens in the future\. Maps mixes today's speeds with patterns from past days, like rush hour\. Busy neighbouring roads are grouped into supersegments, and a neural network \(a program that learns from examples\) lets each one learn from its neighbours\. This work with Google DeepMind made arrival\-time predictions up to 50% more accurate in cities such as Taichung\. ### 7. Picking your route, live Maps compares several possible routes and ranks them by predicted arrival time\. Press the button to cause a slowdown on the main avenue\. Phones there report slower speeds, the road turns red, and Maps switches you to a faster route\. The blue line is the fastest route right now; grey lines are slower choices\. ### 8. Your trip, in real time Position from the sky, a map built from photos and data, speeds shared by millions of phones, and predictions of what comes next all update together\. Your arrival time is a smart forecast, not a promise, so Maps keeps checking as you go\. Before this neural\-network work, Maps' arrival times were already accurate for over 97% of trips\.