Re: [Talk-GB] Import of UK SSSI data

[David Woolley]

On 17/11/2019 22:37, Henry Bush wrote:
I am aware that SSSIs change, so my plan was that my bot would look for 
an existing entry first, and if it exists, either modify or delete it 
(if the latter, I'd verify the tags were the same or something first)


Delete and re-add is something that should not be done, as it destroys 
the history of the object.  Generally if you detect a duplicate, you 
should manually check it before doing anything.


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

Re: [Talk-GB] Import of UK SSSI data

[Henry Bush]

Thank you for everyone's input, I really appreciate it. This is an
interesting project, but it probably doesn't make sense to take it any
further unless it is going to be feasible and the outcome useful.

Responses to some of Jerry's points:

 * The fact that some are mapped makes me a little sad, as it means that my
searches for "site of special scientific interest" and "sssi" on
openstreetmap.org didn't find any of them. I couldn't even figure out how
to find them from that interface you linked to (which I hadn't seen before,
thanks for that): anyone got suggestions for how I can track them down?
* The protected area wiki indicates that SSSIs should be protection_class
of 4, so that's what I was planning to use
https://wiki.openstreetmap.org/wiki/Tag:boundary%3Dprotected_area
* I am aware that SSSIs change, so my plan was that my bot would look for
an existing entry first, and if it exists, either modify or delete it (if
the latter, I'd verify the tags were the same or something first). Thus it
could be run occasionally (by me or anyone else) to update them. Thus
before I start I would find the existing ones and figure out the best way
to handle each: maybe they could just be modified so that my bot would pick
them up and modify them.
* The reason originally that I wanted to add these to OSM was because I
found a completely un-covered area, and most of it was in a SSSI. I soon
realised that it should be tagged with landuse etc anyway, but still I
think that having this information in OSM would be very valuable. If they
render (I haven't tested that yet), I don't think having them covering
private land should preclude their import: after all, this is information
sourced from the original definition, and if the private land owner
disagrees then they should take it up with NE (I assume).
* I would have thought that SSSIs could be a completely separate area from
other classes of nature reserve, for the very reason that this data has a
clear and precise source, and it is likely to be different from any other
areas that might overlap it. If this is not the case and the SSSI should
(in some cases) be defined by specific tags on other areas, then maybe this
project is not viable.
* From Owen's reply, it sounds like licensing is probably not an issue.
This is certainly good news, as obviously that would knock this proposal on
the head very swiftly. I'll do some research myself, but IANAL, and I doubt
I'll be able to decipher the information. I might email NE and see if they
can give me an answer, but I imagine it would be non-committal.

And Phil's:

* Although the nature reserves may be in OSM, I don't think any are tagged
as SSSIs, which is what I would like to do.
* Cross-border SSSIs are definitely a complicated issue, but one that it is
possible to overcome, I feel. Thank you for bringing it to my attention!

And now I just need to go and put these points in to my wiki page as
potential issues!

Many thanks,


Henry

On Sat, 16 Nov 2019 at 17:56, SK53  wrote:

> A few things:
>
>-  A number of SSSI's are mapped
>
> 
>as many are co-incident with various types of Nature Reserves, although
>sometimes there are minor differences in boundaries. For instance the SSSI
>at Newhouse Farm National Nature Reserve is smaller than the NNR.
>- SSSIs are not nature reserves, so protected area is correct. A
>designation, protect_class etc should be considered.
>-  As some are already mapped, any import would need to detect
>collisions & potentially do some quite complex processing if the SSSI is
>not coincident with the element currently tagged with that information.
>This needs to documented. I note that at least one SSSI lies within another
>on OSM which is possibly inaccurate, or reflects historical change (merging
>of 2 SSSIs).
>-  Document which transforms are used to convert from OSGB
>co-ordinates. I suspect we have 3 potential ones in use EPSG:27700, OSTN02
>and OSTN15, see this
> (lengthy)
>doc from the OS.
>-  What is the purpose of adding these to OSM? If they get rendered
>and show up on private land which is not accessible this may have
>undesirable consequences. For 90% of all my purposes I only want SSSIs as
>an overlay and find using the native data from NE/SNH/NRW either as a
>separate layer in QGIS or as discrete tables in PostGIS is perfectly fine.
>The major gripe is having to get data from 3 separate sources (it would be
>4 if NI ASSIs were available as open data).
>-  Virtually all of NE (and SNH & NRW) data is created against
>MasterMap and therefore contains OSGB material. I think, but cannot be
>certain, that NE obtained the necessary permissions for this data to be
>freely usable. Owen Boswarva who occasionally contributes to the list 

[Talk-GB] weeklyOSM #486 2019-11-05-2019-11-11

[weeklyteam]

The weekly round-up of OSM news, issue # 486,
is now available online in English, giving as always a summary of a lot of 
things happening in the openstreetmap world:

 http://www.weeklyosm.eu/en/archives/12536/

Enjoy! 

Did you know that you can also submit messages for the weeklyOSM? Just log in 
to https://osmbc.openstreetmap.de/login with your OSM account. Read more about 
how to write a post here: 
http://www.weeklyosm.eu/this-news-should-be-in-weeklyosm 

weeklyOSM? 
who: https://wiki.openstreetmap.org/wiki/WeeklyOSM#Available_Languages 
where?: 
https://umap.openstreetmap.fr/en/map/weeklyosm-is-currently-produced-in_56718#2/8.6/108.3
___
Talk-GB mailing list
Talk-GB@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-gb

Re: [Talk-GB] Solar farms from REPD, all checked

[SK53]

Great work Dan.

A couple of the larger solar farms which don't seem to have repd:ids are on
water company land. The one I've looked for a few times is the floating
solar panels  on Godley
Reservoir. I've also been unable to reconcile the panels
 (currently
without a surrounding plant) which are at the Five Farms Water Treatment
Plant, just S of Wrexham Industrial Estate. So I have been wondering if
installations on water company land are different in some way.

It's also worth noting that several of us have continued mapping rooftop
solar in the past 6 weeks and we should pass 120k mapped installations in
the next day or so. A big shout out for gurglypipe who has mapped a huge
number around Lancaster as areas. My own focus started in Wrexham district
as I visited it in the middle of October which meant I ground-truthed a
number of ones already mapped, and enabled me to add quite a few others.
More recently I've been aiming to connect existing hotspots by bringing LAs
which connect them up to 50% (Halton, Rochdale & Oldham already done,
Warrington, Barnsley, Kirklees & Doncaster current targets).

There's still a lot of work to ensure every rooftop installation has an
building underneath. Gregory's site reports these at the detailed level,
but not at LSOA or LA. I suppose there's scope to create MapRoulette
challenges.

On the debit side I have also created a lot of fixmes which I can now clear
up: there's an estate on the edge of Runcorn with building integrated solar
panels which looked a bit like solar hot water. Fortunately after having
mapped them I discovered a promotional write up by the manufacturer. I just
need to decide on the relevant tag (see the wiki page

).

Jerry

On Sun, 17 Nov 2019 at 11:36, Dan S  wrote:

> Hi all,
>
> This weekend I completed the task of going through the REPD dataset
> (UK planning data) for solar farms. Out of 1058 entries, about 65 were
> un-spottable,* the rest are in OSM.
>
> We have 908 solar farm objects in OSM for the UK.
> The REPD list totals about 8.1 GW and we've got approx 6.9 GW
> explicitly tagged. (Plus 70 of the 908 have no capacity tagged.)
>
> Most of the solar farms are in there as power=plant. However, there
> are plenty that had previously been tagged as power=generator, and I
> chose not to coerce everything into fixed format. Also, I generally
> didn't trace the panels (nor even the blocks of panels) within solar
> farms, I merely drew outlines. So there's plenty of scope to improve
> the mapping in future!
>
> The repd:id tag is really useful for checking back against REPD. Some
> of the mapped solar farms have multiple IDs (semicolon separated),
> since there are lots of solar farms which had extensions added in
> later years.
>
> As I said, there were about 65 REPD items I couldn't spot in aerials.
> Funnily enough, we have a rather similar number (55) of solar farm
> objects in OSM which have not been associated with a REPD entry nor do
> we have any capacity tagged for them. (Here's a query for non-repd
> solar farms: )
>
> Best
> Dan
>
> P.S. my spreadsheet is still messy, but I updated it as I went, so
> fwiw: http://mcld.co.uk/tmp/wiki_repd_list_dan.ods
>
> * Here are 87 REPD IDs which were either "not seen", or unsure and
> could do with a second eye:
> 1098
> 1176
> 1233
> 1304
> 1325
> 1332
> 1494
> 1515
> 1546
> 1550
> 1587
> 1611
> 1620
> 1716
> 1746
> 1817
> 1827
> 1838
> 1840
> 1900
> 1908
> 1914
> 1975
> 1981
> 2013
> 2015
> 2027
> 2044
> 2060
> 2075
> 2082
> 2089
> 2104
> 2176
> 2204
> 2237
> 2252
> 2274
> 2324
> 2364
> 4713
> 4740
> 4844
> 4857
> 4861
> 4874
> 4884
> 4896
> 5006
> 5063
> 5093
> 5149
> 5152
> 5164
> 5190
> 5232
> 5255
> 5265
> 5320
> 5322
> 5330
> 5360
> 5398
> 5412
> 5440
> 5443
> 5450
> 5472
> 5485
> 5499
> 5506
> 5512
> 5525
> 5543
> 5559
> 5593
> 5603
> 5631
> 5650
> 5793
> 5891
> 5945
> 5977
> 6007
> 6019
> 6108
> 6328
>
> ___
> 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] Solar farms from REPD, all checked

[Dan S]

Hi all,

This weekend I completed the task of going through the REPD dataset
(UK planning data) for solar farms. Out of 1058 entries, about 65 were
un-spottable,* the rest are in OSM.

We have 908 solar farm objects in OSM for the UK.
The REPD list totals about 8.1 GW and we've got approx 6.9 GW
explicitly tagged. (Plus 70 of the 908 have no capacity tagged.)

Most of the solar farms are in there as power=plant. However, there
are plenty that had previously been tagged as power=generator, and I
chose not to coerce everything into fixed format. Also, I generally
didn't trace the panels (nor even the blocks of panels) within solar
farms, I merely drew outlines. So there's plenty of scope to improve
the mapping in future!

The repd:id tag is really useful for checking back against REPD. Some
of the mapped solar farms have multiple IDs (semicolon separated),
since there are lots of solar farms which had extensions added in
later years.

As I said, there were about 65 REPD items I couldn't spot in aerials.
Funnily enough, we have a rather similar number (55) of solar farm
objects in OSM which have not been associated with a REPD entry nor do
we have any capacity tagged for them. (Here's a query for non-repd
solar farms: )

Best
Dan

P.S. my spreadsheet is still messy, but I updated it as I went, so
fwiw: http://mcld.co.uk/tmp/wiki_repd_list_dan.ods

* Here are 87 REPD IDs which were either "not seen", or unsure and
could do with a second eye:
1098
1176
1233
1304
1325
1332
1494
1515
1546
1550
1587
1611
1620
1716
1746
1817
1827
1838
1840
1900
1908
1914
1975
1981
2013
2015
2027
2044
2060
2075
2082
2089
2104
2176
2204
2237
2252
2274
2324
2364
4713
4740
4844
4857
4861
4874
4884
4896
5006
5063
5093
5149
5152
5164
5190
5232
5255
5265
5320
5322
5330
5360
5398
5412
5440
5443
5450
5472
5485
5499
5506
5512
5525
5543
5559
5593
5603
5631
5650
5793
5891
5945
5977
6007
6019
6108
6328

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