Google today announced a major expansion of its capabilities to warn users in advance of crowded public transit. The feature was originally launched in June 2019 and covered approximately 200 cities worldwide, using user-reported data similar to Waze to discover and predict overcrowded trains and buses. Google is expanding that initial rollout in 200 cities to (eventually) more than 10,000 carriers in 100 countries, according to today’s announcement.
Unsurprisingly, Google Maps’ public transit predictions aren’t just based on direct user-reported data. They use AI models trained on that data.
“We apply world-class anonymization technology and differential privacy techniques to location history data to ensure your data remains private and secure,” the announcement reads, although we should point out that attempts to anonymize location data often fail.
Google says its massive transit crowd reporting lets users know whether their transit line is likely to have “a lot of open seats, reach full capacity, or be somewhere or in-between,” so users can make a more informed decision about whether or not to join. board or wait for another train.” In New York and Sydney, the program runs even more in-depth information, providing live information about crowds “down to the transit car level”. Today’s announcement also notes that in the US public transport is less crowded after the morning rush hour and before the evening rush hour – information that probably won’t surprise anyone who uses public transport regularly.
Maps is also upgrading its Timeline feature to include public transit travel history, letting users know how much time they’ve spent and the distance they’ve traveled via public transit, walking, driving, and more. If you don’t like that feature, or don’t trust Google’s “world-class anonymization technology,” you can disable your Android device’s location history entirely. (Disabling Location History significantly affects ride sharing, weather, and map apps, among other things — and can make some completely useless.)