[Talk-GB] 10 fascinating facts about OSM & OSGB

[Chris Hill]

On 18/04/13 16:40, Oliver Jowett wrote:
> On Thu, Apr 18, 2013 at 3:02 PM, Chris Hill <osm at raggedred.net 
> <mailto:osm at raggedred.net>> wrote:
>
>     On 18/04/13 14:14, Shaun McDonald wrote:
>
>         On 18 Apr 2013, at 13:01, "Dave F." <davefox at madasafish.com
>         <mailto:davefox at madasafish.com>> wrote:
>
>             On 18/04/2013 12:52, Shaun McDonald wrote:
>
>                 Updates are a lot harder to do as you have to deal
>                 with differences
>
>             When you say differences, do you mean within the tags?
>             Does it need to do that, could it not do a simpler find &
>             replace?
>
>         If someone has modified the item in OSM and in NaPTAN then it
>         needs manual intervention as to which is more correct - the
>         one in OSM or the one in NaPTAN.
>
>         Shaun
>
>
>     There are various errors in NaPTAN data. The data was created by
>     different means in different local authorities and to varying
>     standards of quality. The position is not always very accurate,
>     but sometimes the accompanying information can be poor too. Some
>     data are very good, but until you survey it you just don't know.
>     As someone who has manually validated every stop in in a city
>     (~1300) and about a third (~900) of the stops in a large Unitary
>     Authority I can vouch for the differences and the wide variation
>     in quality. It became clear that in the city the quality varied
>     from area to area. I suspect that some areas were surveyed or
>     recorded or checked more diligently than others possibly because a
>     more diligent or competent person did the work there.
>
>
> Can I ask what type of changes you made to the NaPTAN data when 
> manually validating the stops?
>
> I've been meaning to look at exactly this - updating the NaPTAN 
> imported data - for a little while now as it's needed before the TNDS 
> data (which has route / timetable info) can be usefully used.
> My rough plan was to assume that all tags under the naptan: namespace 
> were fair game for updating to match the current NaPTAN import. 
> Whether that's practical will depend on if there have been widespread 
> manual changes to the naptan: data or not.
>
I followed the guidelines in the wiki [1] and use the NOVAM viewer [2] 
to see what remained to be reviewed. Positioning is the biggest dilemma, 
how to be sure the position should be moved? Multiple GPS traces to 
average out the road position helps a lot, but a single GPS position is 
subject to the same error as some of the original surveys. Any naptan:* 
tag is fair game to change if it does not agree with your survey.

[1] 
http://wiki.openstreetmap.org/wiki/NaPTAN/Surveying_and_Merging_NaPTAN_and_OSM_data
[2] 
http://mappa-mercia.org/novam/?zoom=15&lat=53.7299&lon=-0.43507&layers=BT&scheme=hull

-- 
Cheers, Chris
user: chillly