A live sat-nav app for rail passengers?

[fandroid]

It arose in a query in the Fares section. Was there an app that you can use live to guide your rail journey as it progresses, updating connections to fit progress both on your train and connecting ones. The general response was no, and that we have juggle apps a bit to stay in touch with progress and potential alterations to best connections.

So do we think it's possible to have a real-time satnav app for rail passengers that advises on connections as the journey progresses? A bit like Google Maps provides for car journeys.

I can imagine that it shouldn't be too difficult to design one for totally flexible anytime tickets, but that the complications might arise on ones that are restricted to specific TOCs and that peak/offpeak restrictions might prove tricky to get right.

 

[Dr Hoo]

Could you flesh out what you mean by “advises on connections”?

All too often I seem to be doing fine until waiting outside Sheffield, Doncaster, Leeds or Stockport only to see the ‘connection’ that I had hoped to make setting off. Then what happens? Advice on local coffee shops or pubs?

 

[gabrielhj07]

For travel in & around London, I find that Citymapper works well for this purpose.

 

[fandroid]

Could you flesh out what you mean by “advises on connections”?

All too often I seem to be doing fine until waiting outside Sheffield, Doncaster, Leeds or Stockport only to see the ‘connection’ that I had hoped to make setting off. Then what happens? Advice on local coffee shops or pubs?

Coffee shops and pubs would be in the first upgrade! The basic one would cover revised advice on feasible connecting trains.

For travel in & around London, I find that Citymapper works well for this purpose.

Is Citymapper live? ie does it automatically change its routeing advice based upon knowing your current location and updated info on your train's arrival times plus connecting services?

Google Maps gives journey planning advice for public transport, much like the various TOC apps or the TfL app, but none of these update without the user intervening.

There's an obvious problem with cities like London, in that if you are travelling
underground, satnav won't know where you are! If the train reporting data is good enough then presumably the software can know which train you are on and so work out your rough position anyway.

 

[stuu]

Won't Google Maps do this is if you use it in sat nav mode? I.e. set the route up and press go on directions... I have never thought to use it that way but it should work

 

[Bletchleyite]

It arose in a query in the Fares section. Was there an app that you can use live to guide your rail journey as it progresses, updating connections to fit progress both on your train and connecting ones. The general response was no, and that we have juggle apps a bit to stay in touch with progress and potential alterations to best connections.

So do we think it's possible to have a real-time satnav app for rail passengers that advises on connections as the journey progresses? A bit like Google Maps provides for car journeys.

I can imagine that it shouldn't be too difficult to design one for totally flexible anytime tickets, but that the complications might arise on ones that are restricted to specific TOCs and that peak/offpeak restrictions might prove tricky to get right.

This strikes me as something the booking sites should do - see a connection that's going to miss and automatically advise what to do based on the validity of the ticket held, and potentially even (where offered close to departure, e.g. LNER) automatically rebook reservations too.

 

[Dave W]

Is Citymapper live? ie does it automatically change its routeing advice based upon knowing your current location and updated info on your train's arrival times plus connecting services?

Not sure it changes routeing advice, other than to advise of next services on the route you're on. Mostly in London, because of frequencies, it pays to stay on the same route anyway. It does update connections though - it uses minimum connection times and/or walking times to establish what you can connect to (for National Rail services, because of the extra connection time, you can sometimes beat it). It might not update to a totally different route, but multiple bus or train routes leaving from the same place will be listed, if they can get you there.

(An example is if I used Citymapper at Manor House to reach The Salisbury Hotel on Green Lanes - it'd offer all of the 29, 141 and 341 buses in a list of next departures from the northbound bus stop outside the station. It'd also tell you to do the quickest thing - which in the evening rush hour is walk)

Citymapper's available quite widely now, in most of the UK's big cities, plus Dublin - and these are quite wide. Birmingham's edition works pretty well on the far side of the Black Country (Stourbridge, Kingswinford, etc).

 

[fandroid]

Won't Google Maps do this is if you use it in sat nav mode? I.e. set the route up and press go on directions... I have never thought to use it that way but it should work

No. Check it out. If you select for a car they provide a Start button and the route advice changes as you travel. There's no start button for public transport, and you need to manually intervene to get the latest advice when you are already en route. It would struggle somewhat to give you valid advice from Your Location if you were already travelling on a train!

It sounds to me that Citymapper does just what Google Maps does for public transport, and every journey planner created by every city transport undertaking, or other apps like UK Bus Checker which covers far more than just buses.

Booking sites main job is to sell tickets, so I'd only expect more independently minded ones like Trainline or Trainsplit to consider it, as a way of hanging on to customer loyalty.

It's more likely that one of the even more independent sites such as RealTimeTrains might see it as an interesting challenge

 

[jon0844]

Google Maps has a long way to go to offer a comprehensive and reliable navigation system for public transport. As said above, it may have real time data for planning, but doesn't take notice of changes or issues, or even know what service(s) you're using to plan for the next leg or even if you got on the bus/train/tram etc. (Yes it can make assumptions via GPS or other location data, but it's not quite the same as being tracked in your car on a specific road).

It also misses out many options.

As a guide it's better than nothing, but it would probably be very hard to do (hence why even Google struggles) given so many different data inputs and variables, and the need to track so many things, as well as not knowing if you've decided to skip the next bus/train to grab some food or pop to the toilet.

Finally, for all the cost, would it be worthwhile? I am assuming most people wouldn't want to pay for such a service so how would it be funded?

 

[Dave W]

It sounds to me that Citymapper does just what Google Maps does for public transport, and every journey planner created by every city transport undertaking, or other apps like UK Bus Checker which covers far more than just buses.

Not sure I agree with that, but I'll put that to one side.

What exactly do you want it to do that Google doesn't?

 

[zwk500]

This strikes me as something the booking sites should do - see a connection that's going to miss and automatically advise what to do based on the validity of the ticket held, and potentially even (where offered close to departure, e.g. LNER) automatically rebook reservations too.

Difficult to check the multitudes of scenarios in advance. Although clearer advice could be given - 'In the event of disruption, your booked ticket entitles you to travel on the next service operated by Avanti' or 'Your ticket is valid on any service between these stations'.

Personally I use Google Maps, once I know my train is going to be late I put in that station as the origin and then my destination and see what it advises.

 

[stuu]

No. Check it out. If you select for a car they provide a Start button and the route advice changes as you travel. There's no start button for public transport, and you need to manually intervene to get the latest advice when you are already en route. It would struggle somewhat to give you valid advice from Your Location if you were already travelling on a train!

It does have a start button if you set it to include leaving from your house and walking anywhere. But that's a good point that it isn't going to know which service you are actually on

 

[Doctor Fegg]

It's certainly possible, though complex.

I'd start by creating an app targeted towards "expert" users - i.e. the sort of people that read this forum. Essentially just provide a direct interface to the relevant data in one app. You could then build out the UI and the smarter logic to make it more of a consumer-facing product.

 

[zwk500]

It's certainly possible, though complex.

I'd start by creating an app targeted towards "expert" users - i.e. the sort of people that read this forum. Essentially just provide a direct interface to the relevant data in one app. You could then build out the UI and the smarter logic to make it more of a consumer-facing product.

The thing is that expert users already know where to look. To cover the development costs you are going to need a wider audience. The key point would be using a real-time and real-space tracking platform (like Raildar) to work out which service a user is on, then using the data feeds to identify when they will arrive at the next interchange station. At that point it's relatively simple to plug in a journey planner like google maps to show the next trains after the arrival time for the quickest onward journey.

 

[jon0844]

To develop all this would be expensive and need constant updating. Even Google doesn't get all the data in a nice easy format to integrate, having to build filters and interfaces that need constantly updating to keep on working.

All of that means who pays? Do you get bombarded with ads on your journey (and would that even be enough?) or will you have to share every bit of information to keep it going? 'Did you travel on this train today?' 'Was it busy on your train' and then a load more questions? I earn a bit of pocket money answering Google surveys, but most people wouldn't want to do this.

Or you go for a subscription model, like TomTom and others, because of the additional data, and then how many people are likely to pay that? A one off sub for irregular travellers, or monthly or annual?

And, as said, most people who would be the ideal target market for such a complex system are probably the people who can already make their own itineraries and keep track of their journey and look ahead.

I don't wish to sound unduly negative (!) as I'd love such a service to be available, thinking of many people who would benefit and perhaps have more confidence to travel on public transport and not just drive or get a taxi (or not go at all), but it really does seem like if Google can't already do it, it's probably not easy.

 

[Doctor Fegg]

The thing is that expert users already know where to look. To cover the development costs you are going to need a wider audience.

Not necessarily!

There are interesting business models around subscriptions/supporter schemes for dedicated users - that's what I do for my site/app (cycle.travel) and it's now funding my development time very effectively. I run the site on a much smaller budget than the big-budget competitors (Komoot etc.), so I don't need the same number of paying users that they have.

I can 100% imagine the same scheme working for a rail app.

(One of these days I'll make an app out of the New Adlestrop...)

The key point would be using a real-time and real-space tracking platform (like Raildar) to work out which service a user is on

Incidentally, there's a startup working on exactly that at the moment: https://www.signalbox.io

 

[fandroid]

Thanks Doctor Fegg. I was hoping someone with a bit of developer knowledge would contribute.

Difficult to check the multitudes of scenarios in advance. Although clearer advice could be given - 'In the event of disruption, your booked ticket entitles you to travel on the next service operated by Avanti' or 'Your ticket is valid on any service between these stations'.

It's that sort of advice that I see as the ultimate development. Sorting out revised connections, without regard to ticket type, strikes me as probably not too difficult, given the multitude of sites that currently present real-time train information. Attaching ticket validity to it is another layer of complexity, but it's the sort of complexity that easily defeats the non-expert user.

Back in the day, split-ticketing was only for the expert traveller. Now, Trainsplit.com makes it look extremely easy and presents the customer with really clear information about the journey.

Looking far ahead, I can imagine the site would have to be really nimble to pick up the sort of situation where ticket acceptance is relaxed in cases of severe disruption.

 

[zwk500]

Looking far ahead, I can imagine the site would have to be really nimble to pick up the sort of situation where ticket acceptance is relaxed in cases of severe disruption.

It should be sorted out much more simply though, because the railway industry should have a single system where this kind of thing is logged properly (not just noted on the incident report) and then a push notification can be sent to all the registered apps, who can then integrate it into issue itineraries.

 

[lkpridgeon]

It would be possible and it's something I've been meaning to work on outside my day job. My main question would be, if disruption were to occur how should the new legs be chosen?

For a journey request we can easily get 5+ options that gives the same overall arrival time. Do we try and stick as closely to the originally requested itinerary, pick the 1st/random or ask the user to choose a new itinerary?

 

[zwk500]

It would be possible and it's something I've been meaning to work on outside my day job. My main question would be, if disruption were to occur how should the new legs be chosen?

For a journey request we can easily get 5+ options that gives the same overall arrival time. Do we try and stick as closely to the originally requested itinerary, pick the 1st/random or ask the user to choose a new itinerary?

The ideal would be to work out the services least likely to be further delayed, but anticipating that would take a mind-reader not an algorithm.

 

[lkpridgeon]

The ideal would be to work out the services least likely to be further delayed, but anticipating that would take a mind-reader not an algorithm.

I suppose it might be possible to base an assumption on historic running data since last timetable change/30 days

 

[trainmania100]

Doesnt do connections but for the satnav type map thing signalbox.io tells you where a train is on a map and finds the service you're on.

 

[zwk500]

I suppose it might be possible to base an assumption on historic running data since last timetable change/30 days

Each incident is different, it would depend what the delay cause is and how the service is planned to get back on time.

 

[fandroid]

It would be possible and it's something I've been meaning to work on outside my day job. My main question would be, if disruption were to occur how should the new legs be chosen?

For a journey request we can easily get 5+ options that gives the same overall arrival time. Do we try and stick as closely to the originally requested itinerary, pick the 1st/random or ask the user to choose a new itinerary?

Good question. When I hear a serious disruption announcement, I try to get an understanding of the nature and location of the incident. I hope that'll tell me if the disruption is going to be a long one and whether all trains on that line are effectively halted. As a good decision requires quite a lot of knowledge I suspect only a system using AI learning could be of any use!

Perhaps initially the rail-journey sat-nav needs to restrict itself to advising the apparent quickest alternative plus the one with least number of extra changes of train, if they are different. Also, flagging any "official" disruption announcements would give the passenger a clue that they might need to seek help. Twitter is often used by TOCs to announce problems. Perhaps that could be linked in too. Google Maps often offers several alternative driving routes on the sat-nav map with times against each one, plus red spots where there are holdups

 

[lkpridgeon]

Good question. When I hear a serious disruption announcement, I try to get an understanding of the nature and location of the incident. I hope that'll tell me if the disruption is going to be a long one and whether all trains on that line are effectively halted. As a good decision requires quite a lot of knowledge I suspect only a system using AI learning could be of any use!

Perhaps initially the rail-journey sat-nav needs to restrict itself to advising the apparent quickest alternative plus the one with least number of extra changes of train, if they are different. Also, flagging any "official" disruption announcements would give the passenger a clue that they might need to seek help. Twitter is often used by TOCs to announce problems. Perhaps that could be linked in too. Google Maps often offers several alternative driving routes on the sat-nav map with times against each one, plus red spots where there are holdups

My thought for an initial implementation is to do a check every 3 minutes against DARWIN for real-time running information against a saved journey request to see if the journey is still possible. We currently store the 5 most recent journey requests in the FastJP website planner however currently don't expose to the user. We also have the ability to trigger a push notifications if the user consents. So overall not too much work to enable real-time monitoring. The next bit would be, do a check to see if a connection is broken (using minimum connection times), trigger a new journey planning request from both the next interchange station and the next itinerary change point. The results compared for the best timetabled arrival time with sanity checks against live running data removing non-viable options. Replace the subsequent legs in the itinerary with new legs and go back to checking every 3 minutes to see if still valid.

It would technically be possible to get data from Twitter however they like changing their API too much for my liking and I don't think TOCs post in a standard format from day to day let alone from one to another! So it might be hard to use the data and would be best linking off to the relevant TOC's page.

 

[Doctor Fegg]

Twitter appear literally to have turned (part of) their API off today so yes!

 

[gingerheid]

Then what happens? Advice on local coffee shops or pubs?

Genuinely unclear as to whether this is sarcasm or genius, though it gets tricky when it's after midnight and the user clicks the "staff refused to organise taxi because it's not our company" button and you need to suggest hotels / overnight coaches / alternative routes to nearby towns / dossing down in Gatwick / spending the night going round the night tube.

 

[fandroid]

Out of interest I had a chance to try signalbox.io on trains today. It identified a whole group of local trains, so not exactly precise on a moderately busy line, and if incorporated on a sat-nav would require user confirmation as to the correct train. Not a big deal, but not automatic.

Genuinely unclear as to whether this is sarcasm or genius, though it gets tricky when it's after midnight and the user clicks the "staff refused to organise taxi because it's not our company" button and you need to suggest hotels / overnight coaches / alternative routes to nearby towns / dossing down in Gatwick / spending the night going round the night tube.

Anything would be better than the automated announcements on evening trains which say "change here for services to X & Y" when the last trains to those destinations have already departed.

 

[lkpridgeon]

Could you flesh out what you mean by “advises on connections”?

All too often I seem to be doing fine until waiting outside Sheffield, Doncaster, Leeds or Stockport only to see the ‘connection’ that I had hoped to make setting off. Then what happens? Advice on local coffee shops or pubs?

Dutch NS do it! In their journey planner they give you a list of station outlets that will be open at the time

Unfortunately, I don't think we have a data feed for it though

 

[Speed43125]

It's certainly possible, though complex.

I'd start by creating an app targeted towards "expert" users - i.e. the sort of people that read this forum. Essentially just provide a direct interface to the relevant data in one app. You could then build out the UI and the smarter logic to make it more of a consumer-facing product.

I wonder if that's strictly necessary for said 'expert' users?

My first time on the former SR network was a journey from Oxford to Littlehampton via Basingstoke where a delay threatened to scuppere a connection. Queue frantic searches on RTT and Google Maps to try and make the original coastway train into LIT, an app to try and sort this out automatically would have to quite advanced to work out not only which trains would help me make the journey but also judgements on what, if any, efforts might be made to get trains back into their paths* - particularly on 4 track railways.

* I incidentally messed this up, and caught an earlier train that was on time but held to allow the delayed train I was originally meant to catch to go ahead of us on the 2 track section...