[Talk-transit] NaPTAN bus stops

[Peter Miller]

On 21 Aug 2009, at 15:49, Chris Hill wrote:

> I've started to survey the bus stops Thomas imported for Hull. I  
> chose a
> small area to look at to see how it went.
>
>    * Out of 12 stops surveyed, 2 don't exist on the ground.
>    * The codes given as the naptan:AtcoCode are not displayed on the
>      stops' signs, but an abbreviated version is, for example:
>      naptan:AtcoCode=2290YHA01230 is displayed as 29901230.
>    * The naptan:Bearing seems good using 8 cardinal points.
>    * The positions are well placed against the road so far, but the
>      position along the road can be quite inaccurate.  This cannot be
>      random GPS inaccuracy as they would equally be away from the road
>      but more likely some kind of human error.
>    * The whole process will take a lot of work!

I have been making a start on the Suffolk data (Ipswich to be precise).

I think we will find the nature of the errors varies by county. I  
found the following:

The two bus stations had all the right number of stops but they were  
poorly aligned to the service roads so I adjusted their positions. We  
need to have a discussion about how to code the stopping points for  
end-on bus stop bays at some point, but lets leave that for a while.

One of the bus stations has bay letters (Bay A etc), the other uses  
public service numbers, so the stop is known as (12 Stop) or (13/3  
Stop) which seems likely to go wrong when new services are introduced,  
but they are correct as per what is written on the shelter.

The bus stops in one of the bus stations needed bearings added which I  
added using the naptan field. Is that ok? or should we populate a  
standard 'bearing' field which we change and always leave the official  
naptan field alone?

The new hospital mini bus station showed four new bus stops, but there  
are only two. I removed the other two, possibly I should have marked  
them 'not present'.

I found about 6 bus stops that were marked as 'customary' but which  
certainly existed and were 'marked' - one even had a lay-by. I have  
added a highway=bus_stop field and changed the NaPTAN field. About  
half the customary stops existed for real, one almost certainly  
doesn't exist (it is very near to a new bus stop), the rest were quite  
possibly real.

There is a common problem in Ipswich of muddling 'Adjacent' and  
'Opposite', so the stop marked 'Back Hamlet (adj)' is actually  
opposite and vice versa. I have corrected a few of those, again, I  
have used the NaPTAN field which I am not entirely happy about but  
where else should it go?

Some stops say that are adjacent to a road when the road has the same  
name on both sides of the road so it isn't clear from the name if the  
side is correct. This naming is not really very satisfactory but that  
is not our problem for now - I will use the bearing as the main  
indicator of the side of the road.

I have had to tweek the positions of some stops to get the on the  
correct side of the road.

More to do, but I have not found it too time-consuming so far as there  
is no need to go into all the side streets and the town is not that  
large. I will just go into town by different routes each time for now  
and then think about some longer trips. Possibly I am not doing the  
job very well!

I guess I have probably tweeked about 10% of them one way or another,  
normally just fine tuning the position.

I have reported the faults I have found via the professional service  
we run for the DfT for the purpose (http://www.itoworld.com/static/naptan 
). I suggest that the DfT/Traveline might consider making this  
interface available to OSM mapping people to do the same or opening up  
a public version.

It will be useful in time to run a comparison between the current  
NaPTAN and the current OSM and produce reports of where they are  
diverging. We would need the permission of the department to take a  
cut of the NaPTAN data before each run but I think they would be  
supportive.



Regards,


Peter


>
> On the stops that are missing I will remove the highway=bus_stop tag  
> and
> add physically_present=no as per the suggestions from the wiki page.
>
> I have found 1 stop (so far) that is not as described.  The stop is
> described as on Gershwin Road, when it is actually on Sibelius Road
> (based on its code).  The bearing is N when the stop points W.  So how
> much of the Naptan data (if any) should I change?  I can move the stop
> to its actual position, but then the naptan:street will not match the
> street it is on and so on.  Any opinions, or suggestions are welcome.
> If we reach a consensus then I'll amend the wiki page.
>
> Cheers, Chris
>
> _______________________________________________
> Talk-transit mailing list
> Talk-transit at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-transit

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-transit/attachments/20090821/52552f2b/attachment.html>