Commons:Batch uploading/Ordnance Survey

From Wikimedia Commons, the free media repository
Jump to navigation Jump to search

Ordnance Survey OpenData[edit]

I spotted the OS map File:Whitehaven area 1 in 250 000 scale.png being used on the high traffic w:Cumbria shootings article and thought it was a copyright violation. I then looked closer, and realised that w:Ordnance Survey has released several datasets under a free (and Wikimedia Commons compatible) license, as part of their OpenData Initiative.

This is a fantastic resource, and of great use to almost any UK geography article. This could greatly standardise mapping across the UK, and replace many custom one off maps with the recognisable and accessible OS standard. There have already been a few OS OpenData uploads to Commons, and they can be found at Category:Maps from Ordnance Survey. A list of all OpenData products can be found at https://www.ordnancesurvey.co.uk/opendatadownload/products.html, the license can be viewed at http://www.ordnancesurvey.co.uk/oswebsite/opendata/licence/docs/licence.pdf

I've not looked into the datasets themselves, but a lot of them are in TIFF format, which would be an issue. And there's also the issue of how we split the maps into separate files. Even if the source data is in separate files, they may arranged in arbitrary grid squares, which may not be helpful if you'd like to show a map of a town. Is there am easy way to display a matrix of multiple maps on Wikipedia?

Still, I don't think these are big issues. These maps are definitely within Common's scope, and can make an immediate impact across Wikimedia projects. Suitcivil (talk) 22:41, 4 June 2010 (UTC)[reply]

Opinions[edit]

The grid

Great sets. License looks ok. I'll take have a shot at it. I wanted to look into these datasets anyway for the Commons:Batch uploading/Geograph upload. Multichill (talk) 08:04, 5 June 2010 (UTC)[reply]

I've downloaded the tiff files. It's about 20G. I'm going to upload the tiff files and jpg versions of the files. I'm going to make a new license template to reflect the OS license. That will probably all be fine, I'm just wondering what the best way is to get these files categorized. All files should end up under Category:Maps of the United Kingdom. Based on the grid square it's probably possible to find the right subcategory.
Each map should link to the other file version (tiff -> jpg, jpg -> tiff) and it would be very nice if every map links to the squares next to it for easy navigation. Multichill (talk) 11:34, 5 June 2010 (UTC)[reply]
At http://toolserver.org/~multichill/temp/OS_OpenData/ I stored the files. I'm currently converting the tif files to jpg files, this will probably take a while to complete. Multichill (talk) 14:04, 5 June 2010 (UTC)[reply]
As these are images with a limited colour palette, and significant line detailing, wouldn't PNG be preferred over JPG? Suitcivil (talk) 15:58, 5 June 2010 (UTC)[reply]
Just for the record, I've spoken to Multichill on IRC earlier today and asked the same. The PNG thumbnailer doesn't render files over 12.5 M pixels. Also there is some work being done one a TIFF thumbnailer, I dont know what the maximum filesize is for that. –Krinkletalk 16:07, 5 June 2010 (UTC)[reply]

After a short IRC conversation, I'm currently experimenting what the options are regarding SVG. There are vector files available. I'll convert a few to SVG and see how our SVG parser is doing. –Krinkletalk 16:07, 5 June 2010 (UTC)[reply]

SVG will probably be quite hard to create. I'm sticking to the tiff and jpeg files. I created Commons:Batch uploading/Ordnance Survey/Template to be substituted, {{Map tile navigation}} to navigate the tiles and {{OS OpenData}} to be used as license tag. Multichill (talk) 21:47, 5 June 2010 (UTC)[reply]
I'm about to upload the first batch jpg part & tif part. Multichill (talk) 15:43, 6 June 2010 (UTC)[reply]
The first (small) batch is online. Ordnance Survey 1:250 000 Scale Colour Raster map gives a nice overview. In this batch I have 1 map per grid square. I have two more batches, one with 100 images per grid square and one with 400 images per grid square. Feedback would be nice. Multichill (talk) 20:07, 6 June 2010 (UTC)[reply]
I should probably add a category to each batch, what about:
Category:Ordnance Survey 1:250 000 Scale Colour Raster maps
Category:Ordnance Survey Street View maps
Category:Ordnance Survey Vector Map District maps
Should I timestamp the files in the filenames? Multichill (talk) 19:00, 8 June 2010 (UTC)[reply]
I'll just keep on talking to myself. New batches at http://toolserver.org/~multichill/temp/OS_OpenData/OS_Street_View_outputjpg/ and http://toolserver.org/~multichill/temp/OS_OpenData/OS_Street_View_outputtif/ . Probably need some fine tuning. Multichill (talk) 20:10, 8 June 2010 (UTC)[reply]
Assigned to Progress Bot name Category
Multichill Uploaded the first batch OrdnanceSurveyBot Category:Maps of the United Kingdom (and subcats)

Boundary-Line data[edit]

I've uploaded PNG and SVG versions of the parish_region shapefile in the bd_line.zip. The png works ok, but lacks detail; whilst the svg is huge: Inkscape runs out of memory for me when I try to do anything with it, and Mediawiki struggles.

IMO, one significant flaw in the data is that the boundaries for Bristol, Liverpool and Torbay (amongst others) include significant area of sea; distorting the coastline. The one data file in the set that doesn't have that issue is the high_water file, so combining the two would produce more useful maps again. My SVG skills aren't up to that job, and in any case I'd have memory issues.

I'm not totally sure but I believe the contents of the various shapefiles in bd_line.zip are:

  • parish_region - Civil parishes (and equivalent, uploaded)
  • district_borough_unitary_ward_region - Electoral wards in England and Scotland; excluding counties with a unitary council (eg Cornwall and Wiltshire)
  • high_water_polyline - Coastline of England, Wales and Scotland (and outlying islands)
  • westminster_const_region - Parliamentary constituencies for all of England, Wales and Scotland
  • district_borough_unitary_region - Admin districts for England, Wales and Scotland
  • county_electoral_division_region - Electoral divisions for certain counties in England (non-Unitary Authorities)
  • unitary_electoral_division_region - Electoral divisions for Wales, and those parts of England excluded in previous file.
  • scotland_and_wales_const_region - Unsure - Welsh assembly and Scottish parliament constituencies I think.
  • scotland_and_wales_region_region - Electoral regions for Scotland and Wales (for the devolved governments)
  • county_region - English counties, which are not Unitary authorities.
  • greater_london_const_region - The london assembly constituencies.

I planning to do location/locator maps for at least:

  1. Districts, parishes, parliamentary constituencies and wards within counties
  2. Wards within Unitary Authorities.
  3. Scottish Parliament, Welsh Assembly and London Assembly constituencies.

I'll do these as png only; mainly as a result of the coastline issue I mentioned.--Nilfanion (talk) 12:47, 8 June 2010 (UTC)[reply]

Nice that data is used!....but isn't this a bit out of the scope of this page? Multichill (talk) 19:00, 8 June 2010 (UTC)[reply]
Well.. yeah sort of off-topic but if I do make a bunch of location maps (over 10,000 easily) that would be a related batch upload :) Figure this is sensible place to ask for 2nd opinion on that.--Nilfanion (talk) 19:23, 8 June 2010 (UTC)[reply]
Fair enough. Did you see the http://openspace.ordnancesurvey.co.uk/openspace/ on how to reuse their data? Multichill (talk) 19:43, 8 June 2010 (UTC)[reply]

Curious as to what you plan to do with the Vector Map District data. The OS provides raster (tif and jpg) data, but its primarily a vector product which unfortunately is in shapefile format only...--~~