[imagesource: Angela Lang/CNET]

Apparently, Google Maps operates differently in Cape Town, and calculates your arrival time so that you’re at least 15 minutes late for any social event.

Arrive at 7PM for a 7PM dinner booking, for example, and prepare to sit awkwardly on your phone for a while before anybody else arrives.

That lone EFF supporter outside Clicks comes to mind, just with less red, and less destruction of property.

Jokes aside, Google Maps is uncannily good at what it does, though, and now the app has given the public a glimpse behind the curtain at how the AI calculates which route to take, the traffic along said route, and when you’ll arrive (ETA).

Google has also partnered with DeepMind, an Alphabet AI research lab, to improve the accuracy of its traffic prediction capabilities.

Johann Lau, the Product Manager at Google Maps, says that the partnership will increase their current 97% accurate trip prediction stats, using machine learning architecture known as Graph Neural Networks.

This from IOL:

“This technique is what enables Google Maps to better predict whether or not you’ll be affected by a slowdown that may not have even started yet,” Lau said…

To predict what traffic will look like in the near future, Google Maps analyzes historical traffic patterns for roads over time.

“We then combine this database of historical traffic patterns with live traffic conditions, using machine learning to generate predictions based on both sets of data,” Lau said…

“We also look at a number of other factors, like road quality, the size, and directness of a road — a highway is often more efficient than taking a smaller road with multiple stops,” he said.

When routes show heavy traffic, Google Maps will weigh up the predicted time of alternate routes and select which of those presents the most optimal ETA.

Like pretty much every other aspect of day to day living, COVID-19 has caused a shift in how Google looks at traffic calculation, with their data showing a drop of up to 50% when lockdowns were put in place:

To account for this sudden change, Google has updated its models to become more agile — automatically prioritising historical traffic patterns from the last two to four weeks and deprioritising patterns from any time before that.

The predictive traffic models are also a key part of how Google Maps determines driving routes.

Google Maps can’t do it all alone, though, and requires authoritative data from local governments related to aspects like speed limits, tolls, or if certain roads are closed due to construction.

The tech also relies on real-time feedback from users, who can report lane closures, accidents, and so on.

I can already hear the Waze fans crying out from the back about how that’s a more efficient navigating app.

You can read comparisons between the apps here and here, but just know that both are owned by Google, so if that was a consideration, rule it out.

If you are sticking with Google Maps, it’s also worth checking out these six useful tools and tricks to make things run smoother.

I don’t have a dog in this fight, so use whichever app you want, as long as you arrive on time.

7PM doesn’t mean 7:15PM, Cape Town.

[source:iol]