Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
Google Transportation Software Technology

Google Maps Can Now Predict How Crowded Your Bus Or Trail Will Be (techcrunch.com) 41

Google Maps is getting a pair of updates that will offer live traffic delays for busses in the cities where it didn't already provide real-time updates, and another that will tell you how crowded your bus, train, or subway car will be. TechCrunch reports: The latter is perhaps the more interesting of the two, as it represents a new prediction technique Google has been perfecting for over half a year. Starting in October, the company began to ask Google Maps users to rate their journey if they had traveled during peak commuting hours of 6 am to 10 am. Google asked about how many seats were available or if it was standing room only, in order to identify which lines had the highest number of crowdedness reports. Over time, it was able to model this data into a new prediction capability designed to tell transit riders how packed their bus or train would be. It also used this data to create rankings of the most crowded routes and stops around the world.
This discussion has been archived. No new comments can be posted.

Google Maps Can Now Predict How Crowded Your Bus Or Trail Will Be

Comments Filter:
  • When Google can predict when some idiot will be listening to some shitty music over his phone's speaker on my bus, then I will have a use for it.
    • Better yet, some fat ass yacking on their phone so loud that everyone can hear how lazy people are where they work?

      Yes, I take the bus every day. End of the work day is much better. Everyone is tired, they just want to get home for a beer.
    • by AmiMoJo ( 196126 )

      It would be great if they could tell you which cars are the most packed. On Japanese trains you often find the cars around the access points to busy platforms get much more packed than those in other locations.

      I often wonder down to the end of the platform because the end cars are usually less used, but not always.

  • by Anonymous Coward

    Can this also predict how crowded truck stops are? As a trucker, this could be very useful. I'd like to plan my stops at times when the crowds will be minimal. It would allow me to make my runs more efficiently, if I spend less time waiting at truck stops.

    By the way, is APK doing alright? It's been a few days since I've seen him at the Pilot Travel Center off of I-81. This is unusual and I am concerned for his well being.

  • by Anonymous Coward

    I think the ranking presented here is most likely BS:
    1) I commute by train almost every day (in Nagoya) and have only been asked once by Google so far what I thought about my (bus) ride, and that was in Tokyo (I also have lots of commuting experience in Tokyo) (I let Google track my movements)
    2) Maybe it all depends on what you mean by "crowdedness", but I am pretty sure that there are quite a few lines in Tokyo that beat the "#9 Tokyo Nippori-Toneri Liner". Here's an article: https://diamond.jp/articles/-/

    • by jrumney ( 197329 )
      People regularly riding small trains like that probably consider it crowded when someone has to stand for more than 2 stops before a seat opens up. On the Yamanote line it ain't crowded until the compression of the overcoats is enough to propel half a dozen people across the platform when the doors open at the next stop.
      • Google actually asks to select one of: many seats available, a few seats available, some people standing, and packed with people standing. On the busy metro lines, it's probably always the latter during rush hour, so Google probably used the length of the busy trajectory and the duration of the rush 'hour' as an addition metric for ranking.

        It only asks when you're using navigation (i.e. when you're a tourist not familiar with the route), so it would be less sensitive to what regular commuters consider crowd

  • by Tom ( 822 ) on Friday June 28, 2019 @02:32AM (#58838634) Homepage Journal

    So basically it takes an average of the statistical data. Which means that as soon as many people start to use the system, everything will fall to pieces because everyone avoids the time that is marked the most busy, moving to a better, less busy time - except that because of that, the "busy" time now becomes less so and the "good" time becomes the busy one.

    I mean, it's not like we didn't see that same effect already. People already adapt their job starting time to the traffic, if they can. They just don't have many data points, so it scatters. If an authoritative, readily available source would say "drive at 7:23 for optimal commute time", then 7:23 would instantly be the most dense.

    • Which means that as soon as many people start to use the system, everything will fall to pieces because everyone avoids the time that is marked the most busy, moving to a better, less busy time - except that because of that, the "busy" time now becomes less so and the "good" time becomes the busy one.

      Even worse, some unsavory folks will try to game the system:

      "Do not take the bus at the time I take it! It's way too crowded"

      "More space for me!"

    • Google has been doing this for years with highway traffic and I'm not aware of a (say, 2-week) oscillation effect due people responding to weekly updates of traffic predictions.

      Weather and large events are probably more of a problem. Especially automatic anticipation of planned local events such as festivals might be difficult.

  • by Anonymous Coward

    For most places in the US, that's too easy. It's Zero, because no public transit exists there.

    Nice to know Google really needed to resort to high powered data mining to determine that.

  • It is not that difficult, really. The nice thing would be if it could predict anomalous situations, one way or the other - I don't need Google reminding me that buses and trains are going to be packed full during the rush hours: I already know that.
  • Prediction is a strong word for Google Maps given that I've driven the same route for over 2 years and it still cannot accurately predict my arrival time within 20% most days (DC Metro traffic is awesome). I can do that on my own without maps with a trip odometer, no fancy algorithms needed. What it will probably do is tell you what the load is "right now", which can be useful, but it is not prediction. If maps could truly "predict", then it wouldn't need to reroute you for anything other than accidents.
  • The designers of Google Maps must not ever ride the bus. When you get to your destination all you want to do is take care of business and return. Simply reverse your source and destination with one button press and you're on your way. That is unless you answer one of Googles questions which is done within the same instance of Google Maps totally obliterating your return trip. The question should be asked in another instance of an app. I'll never answer another stupid Google Map question ever again.

  • I use Google Maps in Houston for my commute downtown, to choose the best route based on traffic congestion. I almost never end up going the route it tells me to go, and I almost always beat its estimate by 5-10 minutes by choosing a route I know will be better.

    For somebody who isn't familiar with a place, this new feature might be helpful. But there's no substitute for personal experience.

Every nonzero finite dimensional inner product space has an orthonormal basis. It makes sense, when you don't think about it.

Working...