Google Maps VASTLY underestimates bike travel times

[bandit957]

Google Maps vastly - and I do mean vastly, bigly, hugely - underestimates bike travel times.

Today, I bicycled to an event in Cincinnati. Poogle Maps said it would take only 35 minutes, even with a regular bike. But I used an electric bike from the local bikeshare program, and it took 45 minutes.

Granted, a lot of the roads and intersections around here are maliciously designed to be bike-unfriendly. But if it takes 45 minutes with an electric bike, Google shouldn't say it takes only 35 with a regular bike.

[Max Rockatansky]

Don't Google estimates usually lean optimistic anyways?

[wanderer2575]

A more general question, although related:  Does Google Maps take red lights, waiting for traffic to clear before left turns, etc. into account when it estimates travel times?

[bandit957]

Don't Google estimates usually lean optimistic anyways?

For travel by car, I'm not so sure.

I remember the driving times in that map in the back of the Rand McNally Road Atlas back in the '80s was very pessimistic.

[bandit957]

A more general question, although related:  Does Google Maps take red lights, waiting for traffic to clear before left turns, etc. into account when it estimates travel times?

I would think it would have to.

[webny99]

Google's walk times also seem to be fairly aggressive - on the few times I've used it for walking directions, I've been able to keep pace, but only barely, and I'm a relatively fast walker.

[webny99]

A more general question, although related:  Does Google Maps take red lights, waiting for traffic to clear before left turns, etc. into account when it estimates travel times?

I would think it would have to.

Absolutely. Making a string of green lights usually takes a minute or two off the travel time, and I'm sure extra time would be factored in for left turns (especially congested ones) using traffic data.

[US 89]

Google's walk times also seem to be fairly aggressive - on the few times I've used it for walking directions, I've been able to keep pace, but only barely, and I'm a relatively fast walker.

I am also a pretty fast walker and I almost always beat Google's walk times by fairly large margins. Perhaps you aren't as fast as you think you are. 

[hotdogPi]

Google seems to assume a constant 3 mph (20 minutes per mile). I'm right about in line with that. I walk quite a lot, an average of 4 miles per day (more like 6 except with a few 0s when the weather is bad), actually getting places when I walk, and I've timed myself as somewhere between 19 and 20 minutes per mile multiple times using the mile markers on Route 110.

I pass other walkers much more often than they pass me.

[SectorZ]

Google Maps cycling routing is more or less 12 MPH. I've been using it for years to learn different routes away from home, and it's always reliably tied to that number. I don't think it takes into account anything that would slow that down or speed it up.

From where I am at home right now, I just decided to make a completely benign and easy 6 mile route, and one thru tons of lights thru downtown Lowell. Each 6.0 mile route gave a half hour time.

[Edit: I found one thing it will take, topography. Up Kancamagus Pass from Lincoln NH is 1:43, down 0:49 for the same 13 miles. That's a 2100' net change.]

[NWI_Irish96]

Don't Google estimates usually lean optimistic anyways?

For car, they overestimate them for me by quite a bit.

[jeffandnicole]

A more general question, although related:  Does Google Maps take red lights, waiting for traffic to clear before left turns, etc. into account when it estimates travel times?

Somewhat. It tends to include normal occasional stoppages. But if you have an extra long light cycle, or have to sit thru a few light cycles, the tines will be thrown off.

There's one light I go thru that has up to a 4 minute, 20 second red, which Google doesn't take into account. Up to that point though, the other lights have shorter reds, and they generally don't affect Google's expected time of arrival.

[Max Rockatansky]

Google seems to assume a constant 3 mph (20 minutes per mile). I'm right about in line with that. I walk quite a lot, an average of 4 miles per day (more like 6 except with a few 0s when the weather is bad), actually getting places when I walk, and I've timed myself as somewhere between 19 and 20 minutes per mile multiple times using the mile markers on Route 110.

I pass other walkers much more often than they pass me.

3 MPH is what I'm getting for the almost exactly for the two mile walk from my house to the gym.  The estimate specifically is 39 minutes on foot from my house to the gym.  Usually when I've mapped out running routes on Google I usually assume the actual run time is about 40% of the walking estimate, tends to usually play out fairly close.

[bandit957]

Google seems to assume a constant 3 mph (20 minutes per mile).

Then they're underestimating walking times too. I haven't been able to walk that fast since I was 17.

[Max Rockatansky]

Google seems to assume a constant 3 mph (20 minutes per mile).

Then they're underestimating walking times too. I haven't been able to walk that fast since I was 17.

The problem with walking and cycling is that there is a pretty large variance for levels of fitness.  I wouldn't expect someone in poor health to walk or ride as someone who is good physical shape.  How does Google make a reasonable estimate for something like that?...estimate by fitness level perhaps?

[GaryV]

The problem with walking and cycling is that there is a pretty large variance for levels of fitness.  I wouldn’t expect someone in poor health to walk or ride as someone who is good physical shape.  How does Google make a reasonable estimate for something like that?…estimate by fitness level perhaps?

Drivers also have varying speeds, ranging from chronic lead-foot to pokey. How do they make a reasonable estimate for that?

The only options are to go by an average or to keep track of a person's travels and use that data for future estimations.

They also have to consider congestion, at least for vehicle travel (it doesn't affect walking much at all, maybe sometimes for bikes). I find that Google often underestimates the delays for congestion caused by things like accidents or construction.

[vdeane]

Google does use traffic data in their directions, but of course anything that isn't right now is using an average based on past data.  I'm not sure exactly how they do it, but I do know that if you travel exactly the speed limit, it will be an underestimate, at least in the northeast.  My 5-7 mph over seems to be just about right on in this part of the country, at least most of the time.  I'm pretty sure it doesn't take into account lights, but it does take into account the effects the lights have on traffic, so most of the time, that's not too far off, at least for vehicles.  Bike and walk directions, of course, don't have such data backing them up, so the effect is far more pronounced.

[webny99]

Google's walk times also seem to be fairly aggressive - on the few times I've used it for walking directions, I've been able to keep pace, but only barely, and I'm a relatively fast walker.

I am also a pretty fast walker and I almost always beat Google's walk times by fairly large margins. Perhaps you aren't as fast as you think you are. 

Or maybe you're just a ninja.    Much more likely explanation is that I just haven't used it enough to get a good read. I know I'm a fast walker because my experience is quite similar to what 1 noted upthread: I quite commonly pass others on walking trails such as this one, but can't ever remember getting passed by someone else walking (as opposed to running or jogging).

[thspfc]

Don't Google estimates usually lean optimistic anyways?

For car, they overestimate them for me by quite a bit.

Probably because they assume people drive the speed limit.

[NWI_Irish96]

Don't Google estimates usually lean optimistic anyways?

For car, they overestimate them for me by quite a bit.

Probably because they assume people drive the speed limit.

I also think they assume people stop more frequently/for longer periods of time.

[webny99]

Don't Google estimates usually lean optimistic anyways?

For car, they overestimate them for me by quite a bit.

Probably because they assume people drive the speed limit.

I also think they assume people stop more frequently/for longer periods of time.

I don't think it assumes the speed limit, at least not by default. I can make up time, but rarely more than about 10% or so, even while driving significantly more than that over the speed limit. Pretty sure it's some sort of algorithm for average travel time and time is saved by driving faster than average.

[vdeane]

I tested it earlier and it assumes ~70mph on the Thruway, even though the speed limit is only 65.  A trip from Rochester to Albany would take a fairly noticeable amount of time longer if one drove exactly the speed limit.  It doesn't factor in stops, either.

This is one reason why I'm hesitant to change my driving to go exactly the limit - I don't want to lose my ability to accurately estimate travel times on Google Maps.  Of course, I don't want my Rochester-Albany drive to take 20-30 minutes longer, either... and 65 on a rural interstate just feels way too slow.

Of course, I also tested it on I-65 in Indiana (still about 70, which is actually the speed limit there), and on roads out west like I-70 in Kansas and I-80 in Wyoming (a little faster, but not a lot - ~73 mph, even though the speed limit is 75/80).

[jeffandnicole]

Google isn't going to take into consideration, especially on local roads, of the time spent slowing down and speeding up.  Even pulling out of a driveway or parking lot doesn't get taken into consideration, but that's going to instantly set you back roughly 30 seconds give or take.  Stopping at the stop sign at the end of your street, which includes that slowing down then speeding up, will be another 30 seconds roughly.  So you're probably a few thousand feet from your home and already off by a minute.

Highways, even for people that drive the speed limit, will sometimes incur short periods of slowdowns, construction-related reduced limits, etc, not factored in.

The way averages work, it's a lot easier to lose time slowing down than it is gaining time speeding up. 

[Roadgeekteen]

I feel like google slightly overestimates walking times.

[hotdogPi]

One thing I have noticed is that Google won't send you on a 15 minute walk between two transit lines (or from your start point to a transit line) even if that's the fastest way. If there is a way to do it entirely on transit, it will do so as if walking wasn't an option.