Major U.K. Real-Time Train Database Opens Up To More Developers | Hacker News Reader

Major U.K. Real-Time Train Database Opens Up To More Developers

techcrunch.com

97 pointsedward57 comments

Half-baked business idea: Train Dating

Allow people to post a profile, and to say "I am currently on this train and potentially interested in chatting to somebody". You have an audience of bored people with smartphones, you have a ready made safe public place, and if your date is not to your taste you have an automatic cut off time. You haven't wasted an evening on a date, you've just multitasked travelling and dating in a low stakes context.

Bootstrapping it to the point where people have a chance of actually finding somebody else on their train would be challenging, but that's what growth hackers are for, isn't it?

Maybe the focus should be on "meeting" / "chatting" rather than "dating". I'd suggest calling the app "Strangers on a Train", although the murder association might be off-putting for some.

Disclaimer: I am clearly not a businessman.

Edit: If anybody reading this is a businessperson and seriously wants to try to make this work, feel free to get in touch with me about doing the tech side of it.

One risk is, if it doesn't work out, you'll be stuck running into the other person regularly, each day on your commute.

You've described Grindr, but for all sexual orientations and with more appropriate geolocation! This could actually work!

If it takes off, the potential impact on train toilet cubicles could be pretty bad. Though the queue might get more interesting.

Name: Brief encounter.

And if you get lucky, "Briefs encounter"

I had a similar idea to this: However it was more of a daydream about how to escape the misery of so many unhappy people wasting x hours of their lives per week.

I thought it could be cool to have a kind of train/carriage discussion topic that you could browse. The most amusing thing about it is imagining people who didn't know about the app wondering onto the train and being overcome with horror at all the people talking to each other :-)

Except one never speaks to one's fellow occupants in the first class carriage.

Unless you are somehow proposing dinner dates in the buffet car. In which case you've created tindr meets top table and potentially solved the monetization problem as well...

There are a lot of 'match.com' and other dating websites advertised on trains. I guess that their predominance there is for a reason so you just might be onto something.

London is the key market to get right to make it work in the UK. Once you have London then things should be easy.

Maybe you need to take your idea to some place such as the Metro or Evening Standard. They have a similar vested interest in the London commuter thing and they are losing out to gadgets at the moment. A dating app could serve their interests if it also took people to their online news content. Plus they have the million or so readers needed for the app to get traction. A profile need not be as extensive as on a normal dating site as it would be more about time and place. Once the profile is made then, when you browse the rest of the Metro/Evening Standard (waiting for those dates to come in) then they get a good idea of who might want to sell you ads. Great, more targeted advertising - rejoice!!!

One problem is the 'I am going to be stalked as soon as I get off this train' factor. So you could change the rules so that only men have a public profile and any women can browse the available men, drop a 'virtual handkerchief' and allow the app to reveal their profile to those they choose.

I know it is not 'Neighbourhood Watch'/Suzy Lamplugh fear any more in the UK, however, your dating app could allow someone to have their best friend know where they are when they are on a date. If they were going on a random-train-date then this might be useful.

If you watched that Panorama programme on dating websites - and the fake profiles - you will know the dating website business model. This is to have subscribers paying for as long as possible, i.e. not finding dates. If you follow HN you will know about the advertiser funded 'Plenty of Fish' and how much money was to be made there by not having the 'classic' model.

Returning to The Metro, they already have a feature where people can have messages about 'the guy with the whatever on the 185' and so forth. If a similar per-train message board existed, of which extracts were in the Metro, then that could be an idea...

I have always used traintimes.org.uk because of the excellent approach the developer took regarding urls: If I want to know the next trains from Leeds to York I can simply go to the address "traintimes.org.uk/Leeds/York".

The homepage suggests the data is obtained "by screenscraping the information on the official [National Rail Enquiries] site", perhaps bypassing the Darwin Database?

thanks for the link, that's certainly better than thetrainline.com which has been frustrating for years. Btw you should stay in York, a bit nicer than Leeds.

Cool, that looks really useful. The number of times I've had to fumble on thetrainline with patchy reception! Do you know any websites that figure out the cheapest combination of tickets to get you to your destination? (Sometimes buying several tickets for legs of your journey is cheaper than buying one ticket for it.)

http://www.redspottedhanky.com/ is supposed to be quite good for that, though I've never used it.

Never use The Train Line. They charge a booking fee. Just your local train operator, they all use exactly the same system, same database, etc but you won't pay £1 to buy the ticket.

Oh I didn't use them to book, only to check which train I needed to catch.

In that case I just use National Rail, the website is well optimised for mobile.

I wrote my own, and my expenditure on trains has fallen from around ukp9000 pa to around ukp5000 pa. It's actively user hostile, occasionally inaccurate, slow, and cranky. A bit like me. But I'm currently on a journey for which I padi ukp18.00 instead of ukp27.40.

It's worth it for me. Email me for details. I will be slow to reply, but I will reply.

Do you know any websites that figure out the cheapest combination of tickets to get you to your destination?

There's an app for that: https://itunes.apple.com/gb/app/tickety-split/id513845919?mt...

Mobile web version: http://splitticket.moneysavingexpert.com/tool.php

Guide: http://www.moneysavingexpert.com/travel/cheap-train-tickets

Don't they all just work on published walk-up fares? My system also uses advances fares and relevant return and/or single journeys. I've never found an app or web site that uses the full range of options available.

Probably because (as I've discovered) it's hard.

Private Eye often have bon mots about how even the ticket offices don't offer people the cheapest ticket because they can't figure it out.

But when you have to deal with things like (eg, I'm hope I'm remembering this right from someone I know) "EUS-CDF being more expensive than EUS-SOT + SOT-CDF even though both halves are the same train as the EUS-CDF journey", you can see how no-one has a bloody clue.

Don't they all just work on published walk-up fares?

I must confess I've never used the app. I just read about it in the MSE email newsletter.

I grew up in London, and didn't often have reason to take a long distance train. I have a vague childhood memory of discovering that London and England weren't synonymous.

About 5 years ago I worked for the guys [1] that provide the journey planning api behind NRE, the train line and most other UK rail journey planners.

We built a public api that as far as I know is still running. It was pretty usable and the pricing was reasonable.

There was always a sticking point about NRE letting us expose the real time data though - hopefully this has now been resolved.

[1] http://www.silverrailtech.com/journeyplanner

I went to OpenTech 2013 and by far the most packed session was Peter Hicks' (Poggs) talk about open rail data[1]. I worked at Rockshore, who maintain the Network Rail datafeeds, and I remain grateful but astonished at how passionate people are about this data.

[1] http://www.opentech.org.uk/2013/audio/streamC/session4.mp3

The huge London-centric contingent of developers likely have to endure a hour-long train trip twice a day, every day. That's basically a month of their life, every year, spent in trains. Obviously they'd jump at the opportunity to improve their life, if given half the chance.

On a related note, Realtime Trains is an excellent website (and app) for showing current (and historical) train running times: http://realtimetrains.co.uk/

It often shows platforms before they're announced in-station, which can be very handy.

I find it shows the platforms weeks in advance.

Just FYI, none the API doesn't function unless you include the following, hope it saves someone some time!

       default_reasons = {
           0: "leaves on the track",
           1: "vandalism",
           2: "drunk chavs",
           3: "driver being absolutely shitfaced",
           4: "solar flares",
           5: "transport police randomly deciding to stop and search everyone on board",
           6: "you having an interview",
           7: "a possible UFO sighting in Belgium",
           8: "British Rail",
           9: "lack of coffee at the previous station, so the driver has just nipped off to get some",
           10: "leaves on the track",
           11: "flooding within 300km of a railway sleeper"
       }
       delay_all_trains_by_minutes = 5
       if date.today().day % 2 == 1:
           delay_all_trains_by_minutes = random.randrange(1,30,1)
       print 'All trains will be delayed by %d minutes(s) due to %s' % \
             (delay_all_trains_by_minutes, default_reasons[random.randrange(0,len(default_reasons),1)])

Also,

12: "the late arrival of a preceding train",

13: "the train driver is currently en-route in a taxi" (I did not make this one up)

and 14: "all the staff are striking"

I also had, we are delayed due to:

"arriving early at the previous station" & "a Swan on the line" & "a passenger refusing to leave the first class carriage, that only has a standard ticket" (police were called, delayed 25 mins)

Personal favourite, heard at Reading station:

"The 14.18 from Great Malvern is delayed due to leaves on the line. Unfortunately in this case the leaves were still attached to the tree."

You forgot the "wrong sort of snow" reason.

Hah, I've not (yet) seen that one! What is it? Yellow snow?

Edit: O. M. G.

http://www.telegraph.co.uk/topics/weather/4434770/Snow-Brita...

   Do you remember "the wrong kind of snow"? That was how
   British Rail tried to explain away the complete 
   disruption of their services in London and the Southeast 
   on February 7 and 8, 1991.

Code review:

default_reasons doesn't need to be a dict. Use a list.

You don't need random.randrange(). Use random.randint() to select a random integer and random.choice() to select a reason from default_reasons.

Also, use the print function rather than keyword to make this all shiny Python3 :-)

default_reasons seems like an odd choice of name, when there is no other kind of reasons. I'd suggest something like reason_choices or reason_options.

Edit: also I'd make it a tuple rather than either a dict or a list.

Don't make it a tuple. A tuple indicates heterogenous data (a record), while a list indicates a sequence of data.

If nothing else, use lists because the syntax means it stays as a list even if you someday remove all but one option and forget to leave a comma, eg.

  REASONS = [
    'leaves on the line'
  ]

  REASONS = (
    'leaves on the line'
  )

After DDGing a bit it seems like you have a point. I did not know about that semantic difference between tuples and lists in python. It seems like this difference is entirely about the meaning conveyed to a human reader, rather than any technical difference. My guess would be that tuples are generally more efficient, but I have no actual data to back that up with.

If I said this was a deliberate .. uh .. irony on my part implying I'm well suited to work at somewhere like BR or Northern, would you buy it?

Fake edit: TY for the review :)

  reasons = [
      "leaves on the track",
      "vandalism",
      "drunk chavs",
      "driver being absolutely shitfaced",
      "solar flares",
      "transport police randomly deciding to stop and search everyone on board",
      "you having an interview",
      "a possible UFO sighting in Belgium",
      "British Rail",
      "lack of coffee at the previous station, so the driver has just nipped off to get some",
      "leaves on the track",
      "flooding within 300km of a railway sleeper"
  ]
  delay_all_trains_by_minutes = 5
  if date.today().day % 2 == 1:
      delay_all_trains_by_minutes = random.randint(0, 30)
  print('All trains will be delayed by {} minute(s) due to {}'.format(
      delay_all_trains_by_minutes, random.choice(reasons))

Way to go OddBloke :-)

I had the first app on the iOS App Store for live UK train times back in 2008. It is no longer available for legal reasons. I wonder if this means I can rerelease it :)

Good point - also they give ample time for changing, and if you are a last-minuter like me sometimes you can leg it and save 20 minutes (there was one long journey I saved an hour on) .. it would be nice to know about these good-day connections.

I remember an early one just called 'Trains' that had a very similar stylistic look to the built-in Weather app.

Liked it a lot, was sad when it stopped working one day.

That was the one :)

Congrats, it was a great app. I hope you rerelease it. :)

Don't know if Tom (Real Time Trains) reads hacker news, but if you do - congrats on the mention! Obsession clearly pays.

Thank you to this article for introducing me to this amazing app / website. Scotrail's app is slow and buggy and awful, so even if the data isn't NRE, the better app alone wins this.

I had an cool idea for an app: When I need a connecting service from the tube to a mainline, my phone should alert me whether I need to run for the train or if I have time for a coffee when I get close to the station. At the moment, I travel at irregular times and have to bring paper timetables around everywhere. I might even have a go at building this, if I have enough time.

Great stuff! The race to build a map of the UK with trains moving around on it begins in 11 days.

I hope that we'll see an app for the UK with CityMapper levels of awesomeness soon, that would be rad. The National Rail apps built by the various train operators are missing some good stuff.

TrainTimes already has a map (http://traintimes.org.uk/map/#bhm) although I'm not sure of their data source.

I've looked at this before and it's already doable through the network rail api (https://datafeeds.networkrail.co.uk/), which differs from the national rail enquiries api and as far as I can remember (and can see) is free.

Playing around with it, there is room for improvement in the accuracy. Quite a few routes and stations are missing.

Yeah definitely, Edinburgh has acquired a few extra rail bridges! IIRC the network rail movement feed tells you when a train has reaching specific passing points, so there is a very granular level of data available if you want to show movements on the track.

OpenTrainTimes already has something like this: it's a signalling diagram rather than a map (railways work in blocks of track rather than absolute train position) and @poggs uses data from the signalling system to show what block a service is in.

See http://www.opentraintimes.com/maps

Is the gagging clause still in place? ( https://mocko.org.uk/b/2013/02/22/you-cant-use-the-live-uk-t... )

I don't know because I haven't yet seen the terms. I'll be keeping the champagne on ice until I do.

Hopefully we will see the return of MyRailLite

nothing about our trains is "Real Time".

Reply on news.ycombinator.com