google

Why is Google Maps so terrible at predicting travel times?

Every time I start a trip of any significant distance, I cue up google maps, which helpfully provides an estimated arrival time. Every time I finish the trip, inevitably later than google thought I would, I wonder why a company with so much data at its disposal is so atrociously bad at guessing how long I'll be on the road. 

Take a recent trip along major highways on the east coast. I left at 9AM. Predicted travel time: 6 hours, 32 minutes. Actual travel time: 9 hours. 

In contrast, google does a terrific job of predicting the duration of a trip across town, or home from my firehouse when I get off duty at 6AM. Intuitively, this makes sense, but when you really think about it, the duration of longer trips should be even easier to predict. 

Why? For starters, little delays during the course of a long trip will tend to smooth out the overall result. If I'm taking a drive of ten minutes and get stuck at two stoplights, adding two minutes to the trip, I've just skewed the results by 20%. But over hours, a delay at one red light will be offset by a green light elsewhere. The longer the time on the road, the more these small factors will tend to move the result toward an average.

Second, google has years of data on traffic speeds, particularly along major roadways. July 4th northbound Interstate 95 between Richmond Virginia and Washington DC? Google could probably predict travel time within 10 minutes for any time of the day, even allowing for one or two inevitable delays due to fender-benders along the way. It's (probably) impossible to predict major incidents, but aggregated traffic data should yield remarkably accurate results for normal days. And the longer the trip, the more accurate the prediction should be, because there should be less variability in the average number of disruptions. 

I have no idea how google guesstimates its travel times, but it appears to use a bafflingly simple approach of calculating the trip as if you were simultaneously passing through all waypoints at the moment of departure. So if you're departing at 5AM, it assumes you'll be cruising at full speed the entire way, instead of projecting forward - hey, it looks like you'll be passing through crushing gridlock in [major metropolitan area] during rush hour... that might just slow things down by, say, hours. Google should be utilizing its vast stores of data on how long it took other people like you passing through the same area during the same time before, even if no such delays exist at the time of departure. Heck, it could throw in 1) during similar weather conditions, 2) on this day of the week or around this major holiday, even incorporating 3) how you drive, if you're willing to let them harvest your data with such abandon.

All I know is, it's far more annoying to be given foolishly optimistic estimates, which are revised backward as you drive, than it is to be given the bad news up front. Listen up, google competitors, if any of you still exist. If you can give me an estimated time of arrival that doesn't diverge from reality more as the duration of the trip gets longer, I will drop google like a hot, inaccurate potato. 

Hold the children

Google organizes all sorts of public data in formats designed to be useful to nonprofits and others. Consider this playable time lapse of the interrelationship between fertility rate (average births per woman) and lifespan. I once worked in reproductive health, mostly in sub-Saharan Africa, and this correlation (note: not causation) was widely known among people in the field. But I never had the resources to illustrate so clearly how countries move along this continuum. 

Degree of separation challenge

Google has just released a panoply of tools/experiments in machine learning and AI, along with source code to allow people who are so inclined to tinker under the hood. 

Degrees of Separation allows you to select two artworks from Google's image bank and allow the system to seek visual connections that connect them. The path between two works by Georgia O'Keeffe passes through centuries and cultures.

Google Sheepview in the Faroe Islands

Residents (well, the tourism bureau) of the Faroe Islands, an autonomous country halfway between Norway and Iceland, wanted Google streetview to come to their archipelago. So an enterprising resident created "sheepview," attaching a 360 degree camera to the back of an obliging sheep and recording wherever it wandered

Apparently google heard about their efforts and wanted in.


Expect more mapping-related posts in the coming month. I've been interested in cartography since my early days solo hiking (and inevitably, getting lost) in the Virginia hills. A few aspects of that experience find their way into my story The Shape of the World's Skin, which will appear in the upcoming SFFWorld anthology You are Here: Tales of Cartographic Wonders. More about that soon. 

Some unexpected ways that self-driving cars will change America

I've been thinking and writing a bit about the advent of self-driving vehicles and how they will change things like safety and liability for accidents. But driverless vehicles will change more than our chances of dying in a car wreck - they may change how we perceive each other.

This past weekend, my wife and I witnessed another driver make one of the stupidly reckless, me-first maneuvers that are commonplace occurrences around the nation's capital. Nothing out of the ordinary there. When I'm on the job, running with lights and sirens, I see people behave in ways that are so frustrating that it can leave me questioning humanity. Most people try to get out of the way the best they can. But a few cut me off, try to outrun me, or get the jump on everyone else by drafting behind me. Some just sit there, blocking traffic, and take no apparent action that would help an emergency vehicle responding to an emergency. It's not their emergency, after all.

When we witnessed the latest instance of bad driving, my wife said "We need self-driving cars, now." It reminded me of this article from the resident of a town where Google is testing out their self-driving vehicles. By his account, the cars are cautious, obey the law without fail, and give pedestrians a lot of room. Another observation: the cars are so careful that other people have learned you can cut them off in traffic when changing lanes.

People tout the safety and efficiency benefits that automatic cars would bring: you can pack far more of them on to overloaded highways, and route them to minimize commute times for everyone (more about that in another post). It's been predicted that they'll cut traffic accidents by 90%, which will likely reduce the current socially-tolerated roadway carnage of over 36,000 annual deaths.  

But what will they do to our souls? Americans equate cars to freedom, to movement and liberty. We love the road, love our cars... and hate other drivers. Driving to work is like being forced to hang out on the comment boards of youtube, or the more unpleasant corners of reddit: you're packed in with a bunch of anonymous strangers who seem determined to behave in the most stupid or meanly opportunistic ways possible. 

The current popularity of zombie movies & TV shows is no great surprise: we're constant participants in a madcap scramble amidst people who appear to have lost their minds. We're daily survivors of the driving dead. 

As self-driving cars begin to appear on the road, expect them to be harassed, exploited, and abused. They'll be the Google Glasses of the street. Their owners will be mocked, their masculinity (and perhaps humanity) questioned.

Then, something will happen. Enough of the vehicles will be among us that a strange shift will take place. People commuting in a self-driving car, sliding along at a mutually-beneficial pace with other vehicles, will look out of their window and see other people. Not zombies, perhaps, but people. Because they're locked in a regulated traffic pattern, they may pace each other the whole way into the city. The gulf between cars will be narrower than it is now. They'll be neighbors for the whole trip. Someone will roll down a window, maybe, and yell a good-natured insult about whatever sports team's logo is plastered on the rear bumper. Freed from the soul-crushing grind that driving is becoming for many people, they won't default to hatred and alienation. 

Or, because this is the U.S., that other person might take out a gun and begin shooting.

But not necessarily. 

As I write it, this sounds like some kind of utopian vision in which sameness makes us all better people. That kind of thing has been skewered in so many dystopian novels that I wonder why we're so terrified of equality. But that's a question for another day. In the meantime, we're on the verge of a revolutionary change in how we move around, which may revolutionize transportation the way the Internet has altered our handling of information. Could it also make us better people? 

All I need is an automated car and a GPS to steer by

Haven't I mentioned the podcast 99% Invisible before? That was an oversight.

In this episode, they talk about driverless cars and the "automation paradox," which goes something like this:

  1. Humans are weak at performing a dangerous task.
    Example: Navigating with a paper map while driving in unfamiliar places.
  2. We create a form of automation that replaces our imperfect performance
    Drive with a GPS and allow it to perform all the navigation for you.
  3. People become weaker at the task because they increasingly rely on the automated solution.
    When was the last time you used a paper map to locate something?
  4. When automation fails or doesn't perform as expected, humans are more prone to failing at the difficult task.

As part of the discussion of driverless vehicles, this episode touches on a question that's been intriguing me for some time. It has to do with how we handle it when someone is killed by a driverless car.

The vast majority of vehicle accidents are due to human error. Automation will reduce that number, but there will still be collisions, and lives will be lost. Right now, we have a two-tiered means of addressing the responsibility for damage to lives and objects: we hold the individual responsible, and we distribute some of the financial liability throughout all drivers in the form of insurance.

But what happens when the fault is in the software, and virtually all cars are running on a single platform (Google)? Will the company take on the cumulative liability of the fatalities caused by errors in their product? Or do we establish a national fund to dispense payouts in the event of fatalities?

That liability might be too great even for Google to bear, but would a tax-funded system amount to a massive governmental payout to private industry? And would the industry actually have less incentive to perfect their technology and save lives, since they would be insulated from liability? 

Anyway, 99% Invisible is all about design, and it's pretty great. Listen to the full episode, with lots of supporting video, right here.