How Does Google Maps Work?

Techquickie ·Techquickie ·2019-05-06 · 1,099 words · ~5 min read
Floatplane YouTube

Transcript

JSON SRT VTT 158
0:00 Thanks for watching Techquickie.
0:01 Click the subscribe button,
0:02 then enable notifications with the bell icon
0:05 so you won't miss any future videos.
0:07 It used to be that looking something up on a map
0:09 meant unfolding a huge piece of paper
0:11 you kept in your glove box
0:13 and driving past the turnoff to Des Moines
0:15 in the time it took to find section A53.
0:18 Fast forward a few years, and now thanks to Google Maps,
0:21 you can zoom in so far on nearly any street in the world
0:24 that whether you want a bodega in the Big Apple
0:26 or just a regular apple in Manhattan, Kansas,
0:29 you're not bound to have a problem.
0:31 But this unprecedented level of detail
0:33 didn't happen overnight.
0:35 Google managed to get a foothold
0:36 in the online mapping scene back in the mid-2000s
0:39 when it acquired WhereTo Technologies,
0:42 who had the idea to let people scroll around on a map
0:44 without having to refresh,
0:46 a novel concept that helped set Google Maps apart early on.
0:50 But they needed more than just a pretty interface.
0:52 Fortunately, Google is first and foremost a data company,
0:56 so they often start by working directly
0:58 with national and local government offices
0:59 to determine not only exactly where every road is,
1:03 but how lanes are spaced and allocated on freeways
1:06 to give an accurate base experience.
1:09 Unfortunately, oftentimes the government won't
1:11 or isn't able to provide a completely accurate picture
1:14 of how an area is laid out.
1:16 Oh, you.
1:18 So this is where Google Street View comes in.
1:20 You can learn more about that up here,
1:22 but hold on, isn't Street View just a way
1:26 for users to virtually sightsee?
1:28 The answer, as it turns out, is no.
1:29 Google also uses Street View internally
1:33 in order to double check and tweak
1:35 its existing mapping data.
1:37 As the Street View car takes photos of its surroundings,
1:39 special AI can recognize any signs caught by the camera
1:43 and send them to be analyzed by Google's servers
1:46 to determine not only lane placement
1:48 and traffic restrictions,
1:49 but among other things,
1:51 even the proper addresses of local businesses.
1:54 All of this information is then also used
1:56 to prevent Maps from giving you bad directions by,
1:59 for example, sending you the wrong way
2:01 down a one-way street.
2:03 That would be bad.
2:04 Adding another layer to Maps is Google's access
2:07 to a huge amount of satellite data,
2:09 partly from its acquisition
2:10 of satellite imaging service Keyhole in 2004,
2:13 and partly from government sources.
2:15 They use top-down views of street markers
2:18 to further improve accuracy,
2:19 and they've even got to the point
2:21 where they can use satellite imagery
2:22 to extrapolate the shapes and footprints of buildings,
2:26 not only to give you 3D views in Google Earth,
2:28 which are super cool,
2:29 but also to enable you to see outlines of these buildings
2:32 in the regular Maps view.
2:33 So by intelligently combining data
2:36 from all of these sources,
2:37 Maps can construct quite an accurate picture
2:39 of any given patch of land pretty well,
2:41 but even Google can't completely automate things, yet.
2:46 They also use human employees
2:48 in a project called Atlas to make necessary adjustments.
2:51 One of this team's common tasks
2:53 is to investigate user error reports
2:55 and tweak routes on the map layer
2:57 that clearly don't match up
2:58 with a recent satellite image or street view capture.
3:01 And this kind of user reporting
3:03 reflects how crowdsourcing is another major piston
3:06 in Google Maps' engine.
3:07 And I'm not just talking about the photos you see
3:09 on business listings, either.
3:10 Google relies partially on volunteers
3:12 to map areas it may have limited data on,
3:15 something that's especially important
3:16 in less developed countries,
3:18 where it's harder for Google to send vehicles
3:19 or get accurate mapping data from a government.
3:22 Of course, Google uses other tricks, too,
3:24 to flesh out the experience.
3:26 Once upon a time, they licensed databases
3:28 to dramatically expand how many businesses Maps would show,
3:32 but these days, listings are pulled
3:34 from other public sources and from business owners
3:36 who are eager to take advantage
3:38 of the visibility that Google Maps provides.
3:40 Owners of large facilities have even submitted floor plans
3:43 and granted Google permission to do site surveys
3:45 to help you find your way around.
3:47 This is particularly useful for sports arenas,
3:49 shopping malls, and airports,
3:51 and Google even works with a third-party service
3:53 that keeps track of gas prices
3:55 by looking at credit card data
3:56 and even getting direct data feeds
3:58 from major gas station chains.
4:00 Finally, whether you realize you've signed up for it or not,
4:03 Google also uses you.
4:05 Location services and the GPS on your phone
4:08 helps Google understand, on top of where it is,
4:10 how to get there,
4:11 and how many square feet of patio space is available,
4:14 what times of day your favorite watering hole gets crowded.
4:17 And they don't show you,
4:18 but they probably know the demographic makeup
4:20 of the clientele, too.
4:22 Wow.
4:22 So the real answer to how does Google Maps work
4:25 is that Google simply has so many information gaps,
4:28 and gathering tentacles,
4:29 that if they were the star of a Japanese anime,
4:33 I probably wouldn't watch it at work.
4:34 Are you interested in computer science?
4:36 Then check out Brilliant,
4:38 a problem-solving website
4:39 that teaches you to think like a computer scientist.
4:42 Instead of passively listening to lectures,
4:44 you get to master concepts
4:45 by solving fun and challenging problems,
4:47 and Brilliant provides the tools and framework you need
4:50 to tackle these challenges.
4:52 Brilliant's thought-provoking content,
4:53 based around breaking up big complexities
4:56 into bite-sized, understandable chunks,
4:58 will lead you from curiosity to mastery.
5:01 And you'll be in the company
5:02 of over five and a half million members
5:04 who share your curiosity and love for math and science.
5:07 So what are you waiting for?
5:09 You can support Techquickie and learn more about Brilliant
5:11 by clicking the link in the video description
5:13 or going to brilliant.org slash Techquickie
5:15 and signing up for free.
5:17 The first 200 people to go to the link
5:18 will get 20% off an annual premium subscription,
5:21 so be sure to check it out.
5:23 That's it for this video, guys.
5:24 Thanks for watching.
5:25 Like the video if you liked it.
5:26 Dislike it if you dislike it.
5:27 That's pretty straightforward.
5:29 Check out our other channels
5:30 and comment below with video suggestions if you have those.
5:34 Subscribe if you want to.
5:35 Follow us on social media or don't.
5:38 It's your life.