Earlier this year, before we entered the pandemic, an article appeared in a number of news sources talking about an artist who “hacked” Google Maps . What he really did was trick the system into thinking there was a traffic jam when there was actually no traffic at all, and it did not involve any computer “hacking” skills. Instead, he exploited a loophole of sorts by filling a little red wagon with 99 smartphones, all logged in to Google Maps, and walked slowly down a street. The real time feedback from the phones to Google’s system made it appear as though there were many people (probably on buses) traveling very slowly, and the mapping system marked it off with a dark red line – indicating a traffic jam. The reality was far different; the street was empty except for the artist and his little red wagon.
While to many this is merely a humorous anecdote, this lesson is critically important to the traffic engineering profession. Google’s map system is essentially a black box – sucking in data and producing colored lines on a map – but there is nothing to tell users the quantity, quality, or fidelity of the data being used. In my experience (as I’m sure it is for many others) it is easy to identify where traffic signals are on Google Maps, as they almost always show up as congested intersections – but unless the light is red when you arrive, there is seldom any actual backup or delay. Again, an artifact of how the mapping system works, but is not clearly shown in the map. It also is incredibly dependent on the number of other users providing real-time data to the system. For someone traveling on a low-volume road, it is not uncommon for there to be minutes of time between vehicles, and this could stretch even further between vehicles actively using Google Maps. Google Maps has been a very useful tool for drivers for years already, but as the artist has shown, it can be very misleading as well because of how it operates. While we may believe the data is in real time, there are unknowns: first, there is some delay in upload, aggregation, and processing back out to the map; second, there is no obvious indication regarding the age of the data (or when it was last refreshed); and third, there is no indication as to the raw number (or rate) of the vehicles used to produce the map output. And, then there is the prediction algorithm that sometimes kicks in based on historical data to substitute for a lack of data…but that is a separate issue.
We already have seen the impacts of Google Maps and its ability to dynamically reroute traffic to avoid a backup. For many of us, we have probably enjoyed skipping past a major delay on a “secret” route. Professionally, however, we realize that Google’s system does not necessarily account for the functional classifications of roadways, and may direct significant traffic volumes onto roads not ever intended for it, which can put both unfamiliar drivers and local residents at a significant safety risk. There are also a large number of stories of drivers being routed onto closed or abandoned roads and routes where right-of-way was dedicated but the road has not yet been constructed.
The greatest risk for the traffic engineering profession, however, comes from the temptation to use the colored lines on a map as a substitute for a traffic study. Like water, it is human nature to try and find an easier path. The Google information is essentially free, readily available and many times representative of what most people see. We have had incidents where traffic study results were called into question because the results did not match expectations from what a person saw on Google Maps. We have even been asked why a Google Maps output cannot just be the substitute for a study.