To predict what traffic will look like in the near future, Google Maps analyses historical traffic patterns for roads over time. For example, one pattern may show that on the 280 highway in northern California, vehicles typically travel at 65 km/h between 6 and 7 in the morning, but only at 15-20 mph in the late afternoon. Google Maps can show you traffic at different times along the route of your choice, based on aggregated location data and historical traffic patterns. As a result, you can plan your trip based on when traffic is easiest.
Google software then combines its database of historical traffic patterns with live traffic conditions. However, to predict what traffic will look like in a few hours, Google needs to analyze historical traffic patterns for roads over time. Using historical data collected over the years and mobile device traffic data using the Google Maps app, the company can create traffic forecasting models for different time periods. Data from these sensors can be used to provide real-time traffic updates. Once collected, the information becomes part of the pool of historical data that is used to forecast traffic on future dates.
In the last 13 years, when you started a route on Google Maps, Google delivered an estimated time of arrival based on years, a combination of live traffic data and historical traffic patterns that was accurate for over 97% of worldwide travel. Early versions of Google Maps relied only on data from traffic sensors, most of which were installed by government transportation agencies or private companies that specialize in compiling traffic data. If you want to see general traffic information, you can use the traffic overlay feature in your Google Maps app. We then combine this database with historical traffic patterns with live traffic conditions and use machine learning to generate predictions based on both data sets.
As more drivers use the app, traffic forecasts are becoming more reliable as Google Maps can look at the average speed of cars on the same route without misinterpreting someone's morning coffee stop as a traffic jam. Since the outbreak of Covid-19, Google has had to deprioritize older traffic data and has changed its models to initially prioritize traffic patterns over the past two to four weeks.
