[iRail] train statistics

Tue Jan 25 19:22:42 UTC 2011

By the way, the NMBS has just given his official feedback for the delay.
The new is only in french, sorry: 
http://www.lesoir.be/actualite/belgique/2011-01-25/ponctualite-des-trains-le-pire-resultat-depuis-1998-817452.php

    * Generally 80-90% train on time depending on the time of year (rush
      hour, holidays, etc..)
    * Who is guilty in delay (46% NMBS - 34% third partie - 18% Infrabel)
    * 25.192 trains canceled = 1,9% total trains
    * 30% of delay caused by hardware issues, 15% third partie issues? -
      11% signaletic issue.

But the interesting fact is that 2 people of NMBS and ACTP (Association 
des Clients des Transports Publics)  are doing a public chat session 
tomorrow noon.
http://www.lesoir.be/actualite/belgique/2011-01-25/chat-la-ponctualite-des-trains-un-probleme-chronique-817480.php

That might be interesting to have a public conversation with them in 
front of newspaper!

What do you think?

Christophe

> 	jonathan orban <mailto:jonathan_orban at hotmail.com>
> 25 janvier 2011 12:03
>
>
> Hi Pieter and all,
>
>
> > Yes, that's one of our plans for February. You summed up exactly what we
> > are going to do. The timeframe that I see is that we should go 'alpha'
> > in March. We will provide historical data through the same API and same
> > calls. At this moment when you search for < +-2 weeks it will generate
> > an error message.
>
> That's great!
>
> > Are you interested in helping?
>
> Given my php and java skills, I'm not going to be of great help for 
> the api. But I'm most interested in designing statistical tools which 
> make sense.
>
> > It needs a lot of thinking: what needs to be saved into a database? Only
> > the realtime information of that day? And does that need to be saved per
> > hour/per minute or per vehicle that day? We don't need to save the
> > static data for each day since that data doesn't change every year. We
> > do however need to take into account that this static data changes in
> > 2013 and has already changed a couple of times in the past;
> >
> > This is an engineering problem for February and I'm sure we will figure
> > this out. As we have already written the real-time API, gathering the
> > information itself is not difficult at al.
>
> This is all very interesting but, unfortunately, I probably won't be 
> able to make it on Feb 2nd. Do you plan on discussing this on the wiki 
> or only in the meetup?
>
> Cheers,
>
> Jonathan
> _______________________________________________
> iRail mailing list
> iRail at list.irail.be
> http://lists.rootspirit.com/mailman/listinfo/irail
> ------------------------------------------------------------------------
>
> 	jonathan orban <mailto:jonathan_orban at hotmail.com>
> 24 janvier 2011 16:46
>
>
> Hi all,
>
> First of all, thanks for the service. I have just started to use 
> irail.be and I find your UI much more simple and fast than railtime.be.
>
> I have a suggestion for the future. I looked at the archives section 
> of the mailing list and didn't find anyone speaking about it so I'm 
> sorry if this is just something that you guys already discussed 
> billions of times, but I thought it was worth discussing it...
>
> What about recording all the data generated by infrabel? I would see 3 
> direct applications that would be very valuable for all travelers:
> 1) Statistics for each train: Because I take the train everyday, I 
> know that the 8h02am IC train between Nivelles/Nijvel and Brussels is 
> almost always on time. But the 8h28 IC train between the same stations 
> is almost always 10 minutes late. It took me about 2 weeks to figure 
> that out. Casual travelers should have access to such info to be able 
> to select between two trains in advance. After discussion with 
> colleagues and fellow commuters, it seems like everyone has his little 
> recipe to get to work on time... What about making this public and 
> based on facts?
> 2) Automatic completion of the "compensation for delays" form: There 
> are very few people who apply for compensations for frequent delays. 
> This is mostly because of the hastle to fill in the form. If it were 
> automatic, plenty of people would start filling in the form. More 
> people applying for compensations would probably urge the nmbs-sncb 
> administrators to improve train punctuality.
> 3) The possibility to verify independently the global punctuality of 
> the sncb-nmbs and evaluate as well the punctuality for specific lines 
> or moments of the day. The official numbers are that about 87% of the 
> trains arrive on time but a study from test-achats/aankop taking 
> numbers from commuters (and hence for peak hour trains) gave a rate of 
> 65%... What is the punctuality rate during peak hours? Publishing that 
> kind of statistics would also probably urge the nmbs-sncb to improve 
> train punctuality...
>
> Questions:
> Is it allowed to record such data?
> How difficult would that be technically?
> How much data would that generate per day?
> How much would it cost to record the data and save it on a server?
>
> Cheers,
>
> Jonathan
> _______________________________________________
> iRail mailing list
> iRail at list.irail.be
> http://lists.rootspirit.com/mailman/listinfo/irail
> ------------------------------------------------------------------------

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/irail/attachments/20110125/4a493148/attachment-0003.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: compose-unknown-contact.jpg
Type: image/jpeg
Size: 1421 bytes
Desc: not available
URL: <http://lists.okfn.org/pipermail/irail/attachments/20110125/4a493148/attachment-0003.jpg>