|
asdf32 posted:Yeah it’s pretty reasonable to not develop algorithms to combat dudes with red wagons filled with 10’s of thousands of dollars of phones. Paradoxish does raise a good point about how thirty people on slow bus does not imply the same traffic conditions as thirty people in slow cars. Google’s understanding of traffic conditions would improve if they recognised “Those thirty phones seem to all be in the same vehicle. Let’s weight the data accordingly.”
|
#
?
Feb 4, 2020 01:14
|
|
|
|
| # ? Jun 7, 2024 23:26 |
|
|
Platystemon posted:Paradoxish does raise a good point about how thirty people on slow bus does not imply the same traffic conditions as thirty people in slow cars. Asking since I have no idea: how much resolution can you actually get on consumer-grade GPS devices? My iPhone gets it accurate down to about 100-200 feet, but will still pretty regularly think I'm standing in front of the apartment building two doors down. Can they pinpoint locations down enough to judge whether those phones are all on one bus, or is it basically "eh, they're roughly around that block or so."
|
|
#
?
Feb 4, 2020 03:03
|
|
|
Sundae posted:Asking since I have no idea: how much resolution can you actually get on consumer-grade GPS devices? My iPhone gets it accurate down to about 100-200 feet, but will still pretty regularly think I'm standing in front of the apartment building two doors down. Can they pinpoint locations down enough to judge whether those phones are all on one bus, or is it basically "eh, they're roughly around that block or so." I think gmaps also uses wifi signals it picks up as a way to narrow down your location further in urban areas. The geolocation API lets you feed it wifi points as well as cell towers and returns lat/long.
|
|
#
?
Feb 4, 2020 03:12
|
|
|
Sundae posted:Asking since I have no idea: how much resolution can you actually get on consumer-grade GPS devices? My iPhone gets it accurate down to about 100-200 feet, but will still pretty regularly think I'm standing in front of the apartment building two doors down. Can they pinpoint locations down enough to judge whether those phones are all on one bus, or is it basically "eh, they're roughly around that block or so." AGPS/whatever it is now is sub meter accurate. You're probably not getting that in a phone, but plenty of high end devices are available that will do it. I was getting this accuracy well over a decade ago with a then-$2500 surveying device running Windows CE lol. (Fire department/hydrant map etc survey being dumped into ArcGIS)
|
|
#
?
Feb 4, 2020 03:12
|
|
|
Sundae posted:Asking since I have no idea: how much resolution can you actually get on consumer-grade GPS devices? My iPhone gets it accurate down to about 100-200 feet, but will still pretty regularly think I'm standing in front of the apartment building two doors down. Can they pinpoint locations down enough to judge whether those phones are all on one bus, or is it basically "eh, they're roughly around that block or so." really good these days, i haven't looked at raw gps data in over a decade but back then it was pretty good but got squirrely in urban areas because one of the things that can cause interference in properly receiving a gps signal is something that causes the signal to 'reflect' like tall buildings, thick forest cover etc. if you're standing in a relatively open area though then even cheap gps can get to within a few feet of your true position from the very helpful website gps.gov 
|
|
#
?
Feb 4, 2020 03:27
|
|
|
A phone GPSr couldn’t reliably locate a person to within a seat width, but there are a couple of few things that would really help in determining whether or not a bunch of phones are on a bus. For one thing, GPS error is largely correlated, within time and location. If you plot each phone’s position on a map, they might all be ten metres from their true ground position, but they’ll all be just about ten metres off in almost the same direction. Phones in neighbouring seats won’t randomly be ten metres from each other. What’s even better is looking at how the phones are moving. Their positions can be queried multiple times per minute, and they also have drat good accelerometers. If all the phones are moving together as a rigid body, they’re in the same vehicle. If not, they aren’t, (or they’re in a bendy bus). Google does give transit directions. They want to know where the buses are so they can give better information to their users, stuff like “this bus is currently running ten minutes late, and there’s a crash ahead that might push that to twenty”. I don’t know how what Google’s current source of bus position data is. Maybe they rely heavily on the transit agency’s own data, but it would make a lot of sense for them to tap into riders’ phones. Maybe they only combine data for scheduled bus services that they know about, or maybe they never actually bothered to treat all bus riders as one entity for traffic calculations.
|
|
#
?
Feb 4, 2020 03:28
|
|
|
Platystemon posted:I don’t know how what Google’s current source of bus position data is. Maybe they rely heavily on the transit agency’s own data, but it would make a lot of sense for them to tap into riders’ phones. transit agencies often publish this data as a public service, because they track their own vehicles for obvious reasons. you can't always rely on someone having a phone being on the bus, especially during off-peak hours when the bus may be nearly vacant
|
|
#
?
Feb 4, 2020 03:30
|
|
|
https://twitter.com/nytimes/status/1224551384365916161 https://twitter.com/leahmcelrath/status/1224582920523931649
|
|
#
?
Feb 4, 2020 16:53
|
|
|
 "No honest man can survive in your democractic party," - Bernie Socrates
|
|
#
?
Feb 4, 2020 17:04
|
|
|
luxury handset posted:transit agencies often publish this data as a public service, because they track their own vehicles for obvious reasons. you can't always rely on someone having a phone being on the bus, especially during off-peak hours when the bus may be nearly vacant For occupancy information (how crowded is a transit vehicle), Google does tap into phone data. Real-time and schedule information is provided by transit agencies in GTFS format.
|
|
#
?
Feb 4, 2020 17:10
|
|
|
Lambert posted:I assume for occupancy information, Google does tap into phone data. Otherwise I don't know how else Google would know how crowded transit vehicles are. most routing apps, including google maps or waze, track user activity because they can use this as a statistical sampling method to get a picture of current congestion levels on roadways. "how full is this bus" is not something which can be accurately derived on a daily basis from checking user locations nor is it very useful information to have for your average bus rider. you can build enough of a dataset over time to predict how full a vehicle might be but this is very much a general guesstimate based on past trends. it's more useful to predict how congested a roadway may be as a driver since you are more likely to route shift to a less crowded route - if you are a transit user, you are much less likely to be able to get a different route to your destination this sampling data is different from the actual physical location of a vehicle or the onboard/offboard count, which would have to be derived and published by the agency themselves from their own equipment rather than passively tracking people's phones and extrapolating that data. most buses these days have cameras on them, you could use this to do image processing to get a very close to accurate idea of how many seats are full at any given time and i'd be surprised if most agencies weren't already doing this in order to track rider and route usage behavior
|
|
#
?
Feb 4, 2020 17:20
|
|
|
Doggles posted:https://twitter.com/nytimes/status/1224551384365916161 What a loving accomplishment.
|
|
#
?
Feb 4, 2020 18:27
|
|
|
Lol, can't wait for tomorrow when we learn actually the app does work, but just randos were logged in.
|
|
#
?
Feb 4, 2020 18:36
|
|
|
I would assume that Google Maps would be using data only from phones actively using GPS navigation with driving directions. no one on a bus would be using Google Maps for active navigation.
|
|
#
?
Feb 5, 2020 00:12
|
|
|
Nah. Your phone is snitching on you all the time.
|
|
#
?
Feb 5, 2020 00:34
|
|
|
luxury handset posted:most routing apps, including google maps or waze, track user activity because they can use this as a statistical sampling method to get a picture of current congestion levels on roadways. "how full is this bus" is not something which can be accurately derived on a daily basis from checking user locations nor is it very useful information to have for your average bus rider. you can build enough of a dataset over time to predict how full a vehicle might be but this is very much a general guesstimate based on past trends. it's more useful to predict how congested a roadway may be as a driver since you are more likely to route shift to a less crowded route - if you are a transit user, you are much less likely to be able to get a different route to your destination Google does present the information how crowded transit vehicles are to the user. And in areas where transit is dense, this very much can be useful information to avoid routes that have high occupancy. And those are tracked through users' phones. I chose less crowded lines with regularity to get to my destination. And yes, counting people boarding/exiting transit vehicles through cameras at the entrances is something agencies do. Laserface posted:I would assume that Google Maps would be using data only from phones actively using GPS navigation with driving directions. They have occupancy information without transit agencies submitting this information through GTFS, so they're absolutely tracking this information through phones. 
Lambert fucked around with this message at 00:47 on Feb 5, 2020 |
|
#
?
Feb 5, 2020 00:41
|
|
|
Lambert posted:Google does present the information how crowded transit vehicles are to the user. And in areas where transit is dense, this very much can be useful information to avoid routes that have high occupancy. And those are tracked through users' phones. I chose less crowded lines with regularity to get to my destination. They get the occupancy data by just plain asking you on your phone how crowded it was when you get off.
|
|
#
?
Feb 5, 2020 00:58
|
|
|
evilweasel posted:They get the occupancy data by just plain asking you on your phone how crowded it was when you get off. Really? I've never been asked this question, and I have their notifications as well as location history turned on.
|
|
#
?
Feb 5, 2020 01:02
|
|
|
Wow, the app really is a tech nightmare. It wasn't deployed through either the Apple or Google stores; it was deployed using a testing platform called TestFairy, a platform that sideloads apps for, you guessed it, testing. Furthermore, the devs didn't even bother to pay TestFairy; they used the free version of the platform that only allows 200 users. https://www.theverge.com/2020/2/4/21122737/iowa-democractic-caucus-voting-app-android-testfairy-screenshots-app-store
|
|
#
?
Feb 5, 2020 01:02
|
|
|
evilweasel posted:They get the occupancy data by just plain asking you on your phone how crowded it was when you get off. this is probably just to fine tune the dataset they're building by looking at trip generation visible to them and then making assumptions about how well that reflects actual total trips like, if you assume you can see 20% of individuals in the local population via passive phone tracking, then you presumably take your dataset and inflate it x5 to get close to the real number. this would be the same process they use for real time or historic roadway traffic
|
|
#
?
Feb 5, 2020 01:04
|
|
|
Arsenic Lupin posted:Wow, the app really is a tech nightmare. It wasn't deployed through either the Apple or Google stores; it was deployed using a testing platform called TestFairy, a platform that sideloads apps for, you guessed it, testing. Furthermore, the devs didn't even bother to pay TestFairy; they used the free version of the platform that only allows 200 users. Dying for Clinton’s app profits.
|
|
#
?
Feb 5, 2020 01:20
|
|
|
Lambert posted:Really? I've never been asked this question, and I have their notifications as well as location history turned on. I get asked frequently. News stories on the feature also explained that was how it worked.
|
|
#
?
Feb 5, 2020 02:20
|
|
|
Platystemon posted:Dying for Clinton’s app profits. Holy poo poo, when is being associated with Hillary's campaign going to turn into the stain it deserves to be?! The Radium Code of election apps.
|
|
#
?
Feb 5, 2020 02:48
|
|
|
FilthyImp posted:Hahahahaha those consultants probably got a several million dollar contract to hire college temps to do all the work. https://twitter.com/kf/status/1224768660138446848 e: It gets worse https://twitter.com/jlofton42/status/1224856054611181568 Arsenic Lupin fucked around with this message at 02:58 on Feb 5, 2020 |
|
#
?
Feb 5, 2020 02:54
|
|
|
Sounds pretty agile to me.
|
|
#
?
Feb 5, 2020 05:00
|
|
|
Laserface posted:I would assume that Google Maps would be using data only from phones actively using GPS navigation with driving directions. What, I use google maps on the bus all the time to figure out how close to my stop I am, or when is the next bus/train coming so I can transfer or maybe I’m in a city I don’t know and want to see how far the walk is when I get off the bus? I’m astounded you think folks traveling by transit wouldn’t be using an app that helps guide you when using transit.
|
|
#
?
Feb 5, 2020 05:29
|
|
|
it's a pretty uncommon use case. one of the nice things about buses and trains is that you don't have to think too much about how to get where you're going. as with all things, your anecdotal experience may vary significantly
|
|
#
?
Feb 5, 2020 05:31
|
|
|
luxury handset posted:it's a pretty uncommon use case. one of the nice things about buses and trains is that you don't have to think too much about how to get where you're going. as with all things, your anecdotal experience may vary significantly That you think this is uncommon is probably why a dude in a wagon just hosed google maps up You don’t have to think about where you’re going but you absolutely have to think about whether or not you’re there.
|
|
#
?
Feb 5, 2020 06:18
|
|
|
Not all agencies yet have a next stop announcement so having an app to tell you to get off is useful, especially when in an unfamiliar area.
|
|
#
?
Feb 5, 2020 07:03
|
|
|
less than three posted:Not all agencies yet have a next stop announcement so having an app to tell you to get off is useful, especially when in an unfamiliar area. what is this hot tech in the US or something, a bus displaying the next stop?
|
|
#
?
Feb 5, 2020 10:47
|
|
|
Goa Tse-tung posted:what In Miami, only a few "important" stops are announced - in between, you either have to look out the window or check with your app.
|
|
#
?
Feb 5, 2020 10:57
|
|
|
less than three posted:Not all agencies yet have a next stop announcement so having an app to tell you to get off is useful, especially when in an unfamiliar area. That's a joke, right? Next stop announcements have been the standard all my life in any city I've been to.
|
|
#
?
Feb 5, 2020 11:38
|
|
|
Antigravitas posted:That's a joke, right? Next stop announcements have been the standard all my life in any city I've been to. He did mention Miami, that place is its own thing because no one wants to admit it's part of the US.
|
|
#
?
Feb 5, 2020 12:26
|
|
|
Regardless the comment was people don’t use maps on public transit and that’s wrong.
|
|
#
?
Feb 5, 2020 12:46
|
|
|
I'm usually listening to music or a podcast on the bus and I'm an absent-minded moron so I find it useful to have a little alert come up when it's time to pull the stop-the-bus lever and get out of my seat, that way I can zone out without paying attention the rest of the time. In my experience a lot of bus riders use Google maps or Transit or apple maps or whatever in the same way I do, although of course it's more prominent among younger people
|
|
#
?
Feb 5, 2020 13:19
|
|
|
uggy posted:Regardless the comment was people don’t use maps on public transit and that’s wrong. Yeah, using Google Maps for public transit is stupidly common. I don't even live in an area well-served by transit and I see it literally every time that I'm on the bus. Why wouldn't you use something that gives you complete end-to-end navigation with walking directions?
|
|
#
?
Feb 5, 2020 15:50
|
|
|
Lambert posted:In Miami, only a few "important" stops are announced - in between, you either have to look out the window or check with your app. alright my bad, also goondolences
|
|
#
?
Feb 5, 2020 15:56
|
|
|
Paradoxish posted:Yeah, using Google Maps for public transit is stupidly common. I don't even live in an area well-served by transit and I see it literally every time that I'm on the bus. Why wouldn't you use something that gives you complete end-to-end navigation with walking directions? i'd say most people who use a transit system or travel on foot use it frequently enough to not need cues as to when to embark/disembark or which path to take. the majority of daily trips tend to be repetitive, especially commutes as the largest individual component of trip category. again, this is just my anecdotal observations, each individual has their own perspective
|
|
#
?
Feb 5, 2020 15:59
|
|
|
luxury handset posted:i'd say most people who use a transit system or travel on foot use it frequently enough to not need cues as to when to embark/disembark or which path to take. the majority of daily trips tend to be repetitive, especially commutes as the largest individual component of trip category. again, this is just my anecdotal observations, each individual has their own perspective You're right that frequent travelers usually know where they need to get on the bus/train and where they need to get off. But they still need some source of information about where they are - in the absence of a phone, the typical source of that information is looking outside, or listening to the announcements from the bus/train speakers. But using a phone to provide that information instead means that you don't have to pay attention to the outside world like that, which means you get to spend more time staring at your phone, listening to your podcast, etc. To be clear, I'm not saying that using your phone is the "correct" way to take public transit - I'm just explaining why so many people do so even when they know the exact route they need to take. Civilized Fishbot fucked around with this message at 16:07 on Feb 5, 2020 |
|
#
?
Feb 5, 2020 16:04
|
|
|
|
| # ? Jun 7, 2024 23:26 |
|
|
luxury handset posted:i'd say most people who use a transit system or travel on foot use it frequently enough to not need cues as to when to embark/disembark or which path to take. the majority of daily trips tend to be repetitive, especially commutes as the largest individual component of trip category. again, this is just my anecdotal observations, each individual has their own perspective well this is my perspective that folks use it even if it's routine - what if there are delays, what if you are trying to connect, what if you can choose different routes based on up to the minute traffic i don't know why you are like fighting folks on this by continuing to say "well this is just what I think"
|
|
#
?
Feb 5, 2020 16:28
|
|