Re: [Talk-GB] Overpass query help - roads within a ward

[Neil Pilgrim]

Hi Tom,

My guess would be that since this wad was added very recently, then the
corresponding 'area', which is updated less frequently (only within
overpass) does not generate results for you. You can see this if you
comment out certain parts of your last query:
- the line showing the relation (A) works fine
- the line generating points around the edges of the relation (A) works fine
- the line generating points within the area (a) derived from the relation
(A) shows no difference, ie. it doesn't work!

Perhaps keeping trying, or just try waiting a day or so and try again :)

--
Neil


On 28 April 2016 at 08:08, Tom Chance <t...@acrewoods.net> wrote:

> Okay, another one with a problem.
>
> Using the exact same query for this ward relation:
> https://www.openstreetmap.org/relation/6173643
>
> Returns an empty dataset:
> http://overpass-turbo.eu/s/fWi
>
> If I run the equivalent query suggesting by Neil, it downloads some of
> ways that are near to the boundary, but not all, and not the ways within
> the boundary:
> http://overpass-turbo.eu/s/fWj
>
> Any suggestions gratefully received!
>
> Tom
>
> web: http://tomchance.org
> twitter: http://twitter.com/tom_chance
> facebook: http://www.facebook.com/TomChanceGP
>
> On 25 April 2016 at 16:45, Tom Chance <t...@acrewoods.net> wrote:
>
>> Thanks Neil,
>>
>> I think I'll just have to go with that solution and then delete the ways
>> I don't want.
>>
>> Tom
>>
>> web: http://tomchance.org
>> twitter: http://twitter.com/tom_chance
>> facebook: http://www.facebook.com/TomChanceGP
>>
>> On 23 April 2016 at 16:42, Neil Pilgrim <osm-talk...@kepier.clara.net>
>> wrote:
>>
>>> I've had a play with this since I've been working with overpass turbo a
>>> bit recently; the best I've come up with so far is something like this:
>>>   http://overpass-turbo.eu/s/fPR (wy over-commented ;))
>>> As you'll see, this does include these roads, but does also include
>>> roads leaving the area.
>>> If you un-comment the relation, then that shows the boundary too, which
>>> I found useful for debugging.
>>> This may be able to be simplified/optimised, but I've gone with what
>>> works, and I switched to using a timeout/json as I'm used to that :)
>>> If the ways missing were part of the relation then I think
>>> way(r.A)[highway][name] should select them too, but that's not how the
>>> relation is composed.
>>>
>>> Hope that gives some other ideas, if nothing else :)
>>>
>>> --
>>> Neil
>>>
>>>
>>> On 23 April 2016 at 14:20, Tom Chance <t...@acrewoods.net> wrote:
>>>
>>>> Hi all,
>>>>
>>>> It has been a very long time since I last posted here!
>>>>
>>>> I'm trying to use the Overpass API to extract all the roads within the
>>>> bounds of a relation, in this case a local government ward. Can anyone spot
>>>> the problem in the data?
>>>>
>>>> Here's the example I'm working with:
>>>> http://overpass-turbo.eu/s/fPI
>>>>
>>>> The problem is that some roads aren't being downloaded with the query.
>>>>
>>>> If you follow that link and run the query, you'll see that the top
>>>> section of Belvedere Road, and all of Landsdowne Place, are excluded. There
>>>> are a few others like this on the edges of the area.
>>>>
>>>> I thought it might be because the way used by the boundary relation
>>>> runs down the same nodes as those roads. But just at the top of those
>>>> roads, the A-Road Church Road is downloaded despite being represented in
>>>> the data in the same way.
>>>>
>>>> I tried putting a 'kink' in the boundary way down at the other end of
>>>> Landsdowne Place where it meets Fox Hill so it 'encompasses' a node, but
>>>> that didn't help.
>>>>
>>>> I hope this all makes sense, and maybe somebody can help?
>>>>
>>>> Tom
>>>>
>>>> web: http://tomchance.org
>>>> twitter: http://twitter.com/tom_chance
>>>> facebook: http://www.facebook.com/TomChanceGP
>>>>
>>>> ___
>>>> Talk-GB mailing list
>>>> Talk-GB@openstreetmap.org
>>>> https://lists.openstreetmap.org/listinfo/talk-gb
>>>>
>>>>
>>>
>>> ___
>>> Talk-GB mailing list
>>> Talk-GB@openstreetmap.org
>>> https://lists.openstreetmap.org/listinfo/talk-gb
>>>
>>>
>>
>
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Overpass query help - roads within a ward

[Neil Pilgrim]

I've had a play with this since I've been working with overpass turbo a bit
recently; the best I've come up with so far is something like this:
  http://overpass-turbo.eu/s/fPR (wy over-commented ;))
As you'll see, this does include these roads, but does also include roads
leaving the area.
If you un-comment the relation, then that shows the boundary too, which I
found useful for debugging.
This may be able to be simplified/optimised, but I've gone with what works,
and I switched to using a timeout/json as I'm used to that :)
If the ways missing were part of the relation then I think
way(r.A)[highway][name] should select them too, but that's not how the
relation is composed.

Hope that gives some other ideas, if nothing else :)

--
Neil


On 23 April 2016 at 14:20, Tom Chance  wrote:

> Hi all,
>
> It has been a very long time since I last posted here!
>
> I'm trying to use the Overpass API to extract all the roads within the
> bounds of a relation, in this case a local government ward. Can anyone spot
> the problem in the data?
>
> Here's the example I'm working with:
> http://overpass-turbo.eu/s/fPI
>
> The problem is that some roads aren't being downloaded with the query.
>
> If you follow that link and run the query, you'll see that the top section
> of Belvedere Road, and all of Landsdowne Place, are excluded. There are a
> few others like this on the edges of the area.
>
> I thought it might be because the way used by the boundary relation runs
> down the same nodes as those roads. But just at the top of those roads, the
> A-Road Church Road is downloaded despite being represented in the data in
> the same way.
>
> I tried putting a 'kink' in the boundary way down at the other end of
> Landsdowne Place where it meets Fox Hill so it 'encompasses' a node, but
> that didn't help.
>
> I hope this all makes sense, and maybe somebody can help?
>
> Tom
>
> web: http://tomchance.org
> twitter: http://twitter.com/tom_chance
> facebook: http://www.facebook.com/TomChanceGP
>
> ___
> Talk-GB mailing list
> Talk-GB@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-gb
>
>
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb

Re: [OSM-talk] [Talk-us] Old Aerodromes

[Neil Pilgrim]

Well, while the use of the data on the 'airport related web sites' may or
may not be valid to use, going around deleting private ones seems somewhat
counter-intuitive - why not simply mark them as private, if that
knowledge/data is open? If they do get rendered, I'm sure there are
precedents for marking public/private ones, and rendering accordingly
(parking comes to mind, but maybe also caves?).

Don't get me wrong - I know that a lot of old/stale data can be present
(particularly in US imports?), and have come across 'schools' which haven't
been so for years, for example. This seems a bit too enthusiastic deletion
though.

--
Neil


On 12 April 2016 at 15:29, Martijn van Exel  wrote:

> Hi all,
>
> Thanks for the feedback. I understand that the existence of an small
> airfield can be hard to verify from imagery - but I am also wondering what
> the value of this unverified and stale data is to OSM. If they were mapper
> surveyed nodes to begin with I would perhaps feel the need to be more
> cautious in removing them. I looked at perhaps 30 of them, looking them up
> on various airport related web sites, and ~70% of them were private air
> strips with no public access from air or ground. So those being
> fundamentally unverifiable (unless there is a sign or some structures on
> the ground that would make it so) I would see no problem deleting them.
>
> I like the suggestion for encouraging additional mapping (runways) if
> visible and this is already part of the instruction, let me know if that
> could be clearer.
>
> I am not so concerned with rendering - that’s not what we map for.
>
> Martijn
>
> > On Apr 12, 2016, at 3:40 AM, Christoph Hormann 
> wrote:
> >
> > On Tuesday 12 April 2016, Martijn van Exel wrote:
> >> Hi,
> >>
> >> I was mapping some rural area in the U.S. and noticed, not for the
> >> first time, an aerodrome node in the middle of a field where there is
> >> obviously no airport or airfield.
> >
> > I am not sure here.  For small airfields the aeroway=aerodrome feature
> > is a fairly abstract thing essentially indicating only that this is a
> > place where aircrafts start or land.  This is not generally something
> > that can be reliably determined from imagery.
> >
> > This is also a problem for map rendering - map styles use these features
> > to place labels and icons but these features are generally too
> > ill-defined and undifferentiated to do this properly.
> >
> > The real observable feature of an airfield is the perimeter fence or
> > other form of delineation which then makes it a landuse mapping but
> > this only works for actively maintained airfields with a clearly
> > visible outline.  Otherwise the observable feature of an airfield is
> > the runway - mapping this is much better defined and more useful
> > information-wise than the airfield itself.
> >
> > So the challenge would IMO make more sense if it would encourage mapping
> > runways if they are visible rather than removing an aerodrome based on
> > the fact that it is not visible on imagery.
> >
> > See also here for a different angle on the problems of aeroway=aerodrome
> > as it is currently mapped:
> >
> > https://github.com/gravitystorm/openstreetmap-carto/issues/1143
> >
> > --
> > Christoph Hormann
> > http://www.imagico.de/
> >
> > ___
> > Talk-us mailing list
> > talk...@openstreetmap.org
> > https://lists.openstreetmap.org/listinfo/talk-us
>
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Nominatim weakness

[Neil Pilgrim]

Would it not simply be sufficient to list viewport results at the top, ie
having a 'local' and 'global' list? Possibly in a small 'tree' view, in the
sense of being able to hide 'local' and 'global' matches separately? (eg
little triangles rotating to indicate if they're shown)

Selecting other things seems to be a matter of knowing the search phrases,
enhancing them, or how easily they can be found/taught, and really is a
distinct problem IMO.

--
Neil


On Monday, 14 December 2015, Jorge Gustavo Rocha  wrote:

> Hi,
>
> I think that we can add an option to bound the results to the current
> viewport. That option would be passed to nominatim or any other search
> engine.
>
> Personally, I would prefer the search bounded by default, but users could
> change it to "everywhere" to see additional results.
>
> Regards,
>
> J. Gustavo
>
> Às 11:17 de 14-12-2015, Martin Koppenhoefer escreveu:
>
>>
>> 2015-12-14 10:04 GMT+01:00 Sarah Hoffmann > >:
>>
>> If you type in 'Starbucks' in the search box, then you just get
>> objects named that way. No difference with searching for, say Berlin.
>>
>> Now, if you type 'cafes' in the search box, then you are probably
>> looking for all amenity=cafe and that is a POI search. It's true
>> that this particular query doesn't work on osm.org .
>> You have to
>> additionally specify a place, e.g. 'cafes in Poughkeepsie' actually
>> returns the one Starbucks in town.
>>
>>
>>
>> if I search for
>> "Starbucks in Holland" I get not hit, neither for
>> "Starbucks in the Netherlands"
>> "Starbucks in Netherlands"
>> "Starbucks in Amsterdam"
>>
>> but "starbucks, Netherlands" yields a lot of results (there's even a
>> restaurant among them). Btw., also "starbucks, Holland" leads to a lot
>> of results (not sure if they are the sama, they are in different sorting
>> order at least). You can even omit the comma, but you shouldn't add an
>> "in" because this will lead to no results.
>>
>> Cheers,
>> Martin
>>
>>
>> ___
>> talk mailing list
>> talk@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk
>>
>>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [Talk-GB] Data search request: help please

[Neil Pilgrim]

A quick search on FHRS (which admittedly doesn't do all of them) seems to
support the two-pub hypothesis (excepting maybe a bowling club).

--
Neil

On 22 January 2015 at 22:31, Frederik Ramm frede...@remote.org wrote:

 Hi,

 On 01/22/2015 10:59 PM, Andy Mabbett wrote:
  I want to compile a list of all the pubs (including taverns bars, etc)
  in the UK, with the word Louise as part of the name.

 I, too, found a much smaller number than expected, by querying a
 database I have installed for rendering purposes:

 select 'W' as type,osm_id,coalesce(amenity,tourism) as what,tags-'name'
 as name,st_x(st_centroid(way)) as lon,st_y(st_centroid(way)) as lat from
 planet_osm_polygon where  tags-'name' like '%Louise%' and way 
 st_makebox2d(st_makepoint(-14.02,49.88),st_makepoint(1.83,61.07)) and (
 amenity in('pub','bar','restaurant','fastfood') or tourism='hotel' );
  type |  osm_id   | what |  name   |lon |
 lat

 --+---+--+-++--
  W| 164418353 | pub  | The Bree Louise | -0.135489232160522 |
 51.5270537134717
  W|  97237916 | pub  | Princess Louise |  -0.12157304700441 |
 51.5172546574115
 (2 rows)

 And the same query for points gives zero results.

 Bye
 Frederik

 --
 Frederik Ramm  ##  eMail frede...@remote.org  ##  N49°00'09 E008°23'33

 ___
 Talk-GB mailing list
 Talk-GB@openstreetmap.org
 https://lists.openstreetmap.org/listinfo/talk-gb

___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Upcoming changes to OpenStreetMap.org website

[Neil Pilgrim]

I wondered when the SOTM-suggested redesign would surface :)

For me (regarding the redesigned version):
- The mobile website looks better (menus hidden in the typical menu
section), though the top title bar could be shrunk a little IMO.
- The regular website is better on my little netbook screen
a) the wider/shorter search/welcome piece is an improvement, even if
simply due to the absence of a vertical scroll bar due to the height
of the previous left-hand sidebar.
b) I would second (third?) the comment re the ability to 'X' or drop
down/up the welcome piece, though when to have it (re)display by
default (new window/tab, new session, cookie/duration) isn't clear.
c) Another thing is that the Edit box requires log-in, so... is there
a reason that the 'log in' isn't the main route to editing, or should
it really be greyed out until the user is logged-in (with a
hover-tip)? I suppose this is about making people aware that it can be
edited (online)?
d) Lastly, the same point applies as per the mobile website, that the
top white bar is higher/deeper now; it may be more aesthetically
pleasing, but for those of us with tiny screens it's a little
frustrating?

Do we need to forward our comments to the talk list, or is there some
other official channel to make comments to?

--
Neil


On 16 November 2013 18:01, ael law_ence@ntlworld.com wrote:
 On Sat, Nov 16, 2013 at 01:36:40PM +, David Earl wrote:
 On 15/11/2013 20:15, Rob Nickerson wrote:
 (The aim of this email is to provide prior knowledge of an upcoming change 
 to the
 OSM website and to give you an opportunity to provide constructive feedback)


 Would it be possible to have a dismiss button on the Welcome to
 OpenStreetMap box (including Learn More and Sign Up, but not
 including search)? Just like the x on the panels that replace it,
 e.g. when you search. Perhaps if you dismiss it it could join the
 green buttons as 'Welcome' to get it back.

 On iPad and netbooks, this box takes up a substantial part of the
 screen obscuring the map. Less of a problem on larger screens, but
 still intrusive if you want to see the whole map. Clearly it is a
 very important part of the page when the whole point is to promote
 OSM, but being able to make it go away would be helpful when you're
 just trying to make use of the map.

 +1

 ael


 ___
 Talk-GB mailing list
 Talk-GB@openstreetmap.org
 https://lists.openstreetmap.org/listinfo/talk-gb

___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] UK Food Hygiene Rating System

[Neil Pilgrim]

On 19 Oct 2013 15:15, Andy Street m...@andystreet.me.uk wrote:

 On Sat, 19 Oct 2013 15:02:01 +0100
 Dave F. dave...@madasafish.com wrote:

  Importing much of another database seems a bit pointless to me when
  much of the data can change. I feel just a reference back to it is
  suffice.

 +1
+1

 A quick look at taginfo suggests that fhrs:id would be an appropriate
 way to tag this:

 http://taginfo.openstreetmap.org.uk/keys/fhrs%3Aid

I'm pretty new to what to do with 'designing' new tags, but that isn't many
existing ones really (is it?), and isn't ref more in line with existing
usage? ie fhrs:ref ? or is id in common use in other cases?

  Whether it should be a full URL, or just FHRSID=516821, I'm not sure
  about. Which would be better for web page design/rendering?

 I'd opt for just the id rather than a URL. Database primary keys
 are generally changed infrequently and the URL may well change if the
 FSA decide to redesign their website.

+1

--
Neil
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Ongoing routing develiopments

[Neil Pilgrim]

On 24 June 2013 17:48, Andrew M. Bishop osm-li...@gedanken.org.uk wrote:

 May I suggest that you give Routino (http://www.routino.org/) a try?
 There is a UK example running on that website that you can try out.

 Disclaimer: I wrote Routino, so I might be slightly biased.

Perhaps I missed it, but is there a last updated from OSM date/time
listed somewhere on the test site? just tested it locally and I
thought I added a turn restriction - though admittedly the first I've
done - and it seems to have ignored it.

Thanks,

-- 
Neil

___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

[Talk-GB] (Olympics) Gold postbox map - credits missing? add to metadata on tagged postboxes?

[Neil Pilgrim]

Hi,

I noticed a few days ago that the royal mail has made a map of the
postboxes they've painted gold (just temporarily?) for this Olympics.
Quite nicely, they use OSM, but they don't (obviously? at all?) credit
OSM:
  http://www.goldpostboxes.com/
Is it credited? If not, who follows up these kinds of things?

The post box positions as marked on that map are an overlay of the OSM
map, so could give hints as to positions of untagged postboxes for
later surveying.

I saw an earlier discussion which indicated a lack of enthusiasm (my
impression at least), regarding adding olympics data as it was
considered 'transient' or 'historical' - or at least in a few
weeks/months time it will. What's the feeling regarding tagging the
'gold' nature of these for 2012? Immediate ideas that spring to mind
are something like tagging them as gold temporarily and then adding a
prefix like 'was:' (akin to disused:). Alternatively, since post boxes
are all ref'd (in theory!) then an external database suffices...
except that lots of extra data is already *in* the database...

Thoughts?

-- 
Neil

___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] (Olympics) Gold postbox map - credits missing? add to metadata on tagged postboxes?

[Neil Pilgrim]

On 7 August 2012 17:19, Nick Whitelegg nick.whitel...@solent.ac.uk wrote:

I noticed a few days ago that the royal mail has made a map of the
postboxes they've painted gold (just temporarily?) for this Olympics.
Quite nicely, they use OSM, but they don't (obviously? at all?) credit
OSM:
  http://www.goldpostboxes.com/

 This has come up before on the talk list.
 The site does now credit OSM - look at the bottom.

Ah, ok, thanks. I should probably subscribe to -talk as well as -gb...
though the torch route came up on -gb, which was why I posted here. I
naively expected the credit to be on the map...

There is some hint of the olympics on the wiki, but no obvious tagging
to date, at least by my taginfo searches.

-- 
Neil

___
Talk-GB mailing list
Talk-GB@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-gb