Predicting traffic conditions on the roads and determining the best routes for a journey is an incredibly complex task. And yet, Google Maps does it in a matter of seconds – every single time. So, how does the navigation app figure out whether the traffic along your route is heavy or light or what your estimated time of arrival (ETA) would be?
Getting live traffic estimates — whether or not a traffic jam will affect your drive right now – is pretty straightforward. With more than 1 billion kilometers driven with Google Maps in more than 220 countries and territories around the globe, aggregate location data helps to understand the current traffic conditions on the world’s roads.
But this information does not account for what traffic will look like 10, 20, or even 50 minutes into your journey. And this is where advanced machine learning techniques come into play.
To predict what the future is going to look like, Google Maps takes a peep into the past. Analyzing historical traffic patterns over time, Google has learned what road conditions could look like at any given point of the day. “For example, one pattern may show that the 280 freeway in Northern California typically has vehicles traveling at a speed of 65mph between 6-7 am, but only at 15-20mph in the late afternoon,” explains Johann Lau, Product Manager, Google Maps.