Question 1:

• What is the license of the data itself?
• We would possibly want to filter the data on one or more of the Type / Register / Type of register fields but otherwise I think it might be suitable data

Question 2: Per https://commons.wikimedia.org/wiki/Commons:WLM Australia has never participated in WLM and I couldn't find them in the monuments database. There might be Wikipedia lists which were not imported to the database though.

So at the bottom of e.g. http://www.environment.gov.au/cgi-bin/ahdb/search.pl?mode=place_detail;place_id=18843 I see that the license is "CC BY 3.0 AU" and a link to their copyrights page. From that it doesn't look like the data is free enough to import to Wikidata.

Now since we are not interested in all of the data the question is if there is a subset which can be released and used on Wikidata?

Is all of the data copyrighted at all? How strict is Wikidata about the creativity threshold? The long description and photos are not necessary for the import, I guess. Would it be OK if you'd stick to facts alone?

With regards to copyright the biggest problem is probably issues similar to the database protection rights in Europe.

Per the quote (my highlights) it seems as though Australia has something similar with creativity being interchangeable for sweat-of-the-brow.

A database will be protected under Australian law if it is a literary work; expressed in material form; meets the originality test; and has a relevant connection with Australia. Facts and data in themselves are not protected by copyright. However, a collection of data, a dataset, or a database may be protected by copyright if it is sufficiently original. Whether a work is sufficiently original to be protected by copyright depends on whether it has been produced by the application of independent intellectual effort by the author/s, which may involve the exercise of skill, judgement, or creativity in selecting, presenting, or arranging the information. This summary synthesises recent cases regarding originality in factual compilations.

Met Kerry today, conclusion is to focus only on the federal lists (National Heritage List & Commonwealth Heritage list), those should be CC-BY. So if the copyright is the blocker, I suggest to use the oldfashioned lists for now - and worry about transfer to wikidata later?

commonwealth heritage list spatial data is available here: https://data.gov.au/dataset/57720684-4948-45db-a2c8-37259d531d87

I've attached the coordinates that @Yarl extracted from the National heritage list at

@Lokal_Profil just to confirm: is there anything missing to get this rolling from your end?

Sourcing: So the list that we are basing all of this of is that which was sent from @Gnangarra to @Effeietsanders with 520 entries (P5911). To source the data on Wikidata I need info on where this comes from (Either a dataset with an equivalent item on Wikidata or a url) and when it was downloaded.

The same goes for the coordinates which @Yarl extracted (P5881).

Coordinates for the commonwealth list (P5912) can be sourced as an url but need an item if we want to source it more properly.

Mapping: I've mapped all of the source fields to Wikidata in d:User:André_Costa_(WMSE)/COH/Australia. Please take a look and make sure it makes sense. There are currently two unmapped fields which I'm happy to get input for.

Please also add the sourcing information (above) to this page.

Issues with source data:
Three id numbers (place_id) are repeated in the dataset from @Gnangarra: 105694, 105698 and 105707 (all on the National Heritage list). This needs to be resolved as they are otherwise not unique and cannot be mapped against the relevant coordinate.

In the coordinates dataset created by @Yarl 105891, 105704 and 106065 are used multiple times making it unclear which coordinate to use.

De-duplication:
Other that looking for wikidata entries that already use the P3008 (place id property) are there any other ways of automatically identifying heritage sites already on Wikidata?

• I'll leave the question for Gnangarra for him.
• For the coordinates that Yarl extracted: these came from http://www.environment.gov.au/heritage/places/national-heritage-list and were downloaded on 13 August.

De-duplication could happen partially based on the coordinates, I guess? But this could only give suggestions - and will never be error free. Otherwise, it's probably manual work, I'd guess.

This is excitingly close to working :)

If someone creates a list of manual tasks for muggles to do I would be happy to do some of them.

Commons
In addition to the Wikidata side of things there are also things that need doing on Commons for Australia to be ready.Possibly @Romaine
could help with these?

Specifically the following are needed:

• An upload campaign
• Any localised WLM templates
• Categories for "Listed <something> in Australia"
• A template holding the place_id and categorising these in "Listed <something> in Australia with known IDs" with the id as sort key.

Finding monuments
Finally how are contestants finding which images to upload? Monumental? Listeria powered lists? Old-style lists? For both of the latter two templates are needed on whatever wiki they will live and someone with knowledge of that wiki and its policies will have to decide where they should live.

Replies

In T153221#3548535, @Effeietsanders wrote:

• I'll leave the question for Gnangarra for him.

I'm not sure (s)he is actually on Phabricator which is why I tagged you.

• For the coordinates that Yarl extracted: these came from http://www.environment.gov.au/heritage/places/national-heritage-list and were downloaded on 13 August.

Thanks I've added this to the document.

De-duplication could happen partially based on the coordinates, I guess? But this could only give suggestions - and will never be error free. Otherwise, it's probably manual work, I'd guess.

Coordinate based de-duplication (or suggesting duplicates for a human to review) is probably the best. This will not happen before WLM though (I'm also not saying I'm the best person to build this) so my guess is that extended de-duplication will have to happen afterwards.

Also a ping that there are still two question marks with regards to the mapping. Feedback (on the talk page) is appreciated otherwise I will make an uninformed guess ;)

(some kudos for Romaines supertable: https://commons.wikimedia.org/wiki/User:Romaine/Wiki_Loves_Monuments/2017/table )

• Campaign: https://commons.wikimedia.org/wiki/Campaign:wlm-au
• Category on Commons: seems to exist on https://commons.wikimedia.org/wiki/Category:Heritage_buildings_of_Australia, but this should get subtrees for the specific types of heritage, provinces, etc. This will need specifically two subcategories: one for the National Heritage List, and one for the Commonwealth Heritage List. We could consider building the tree already for the Register of the National Estate, which may be included next year. See http://www.environment.gov.au/heritage/publications/australian-heritage-database .
• For the identifier, I would suggest one single identifier, based on the 'entry in the Australian Heritage Database'. Possibly on top of that, we can create templates for the National Heritage List, Commonwealth Heritage List and register of the National Estate or even places under consideration. This would allow a single template for uploading, which can then automatically be replaced with the more specific template.

I'll poke Gnangarra for this thread.

Item 105694 is Lord Howe Island the first instance(424) is the area of the main island (Q104784) and the second instance(425) appears to be the area of the additional islands in the group which is covered by (Q1869866) and includes Wolf Rock(Q2695142), Sail Rock(no id) and Mutton Bird Island(no id) - combine as one entry under (Q1869866) which is the whole group of islands

Item 105698 - Macquarie Island (Q46650) - same as before multiple Islands though in this case there is no group, thought the first instance appears to be related to Judge and Clerk Islets (Q46439) & Bishop and Clerk Islets (Q46489) both which are part of the Macquarie Island Nature reserve (no WD item)

Item 105707 - combine as Heard Island and McDonald Islands (Q131198), the first instance is McDonald Island (Q1915097) and the second one is Heard Island (there is no WD id)

for 105891, 105704 and 106065

105891 - the Australian Alp NP, the different Coordinates are for individual Peaks within it
105704 - Gondwana Rain forest, different locations within the rain forest
106065 - two ships HMAS Sydney(Q1031986) @ 26°14′31″S 111°12′48″E, Kormoran(Q708046) @ 26°05′46″S 111°04′33″E both sank off Geraldton in WA following after attacking each other in WWII

Thanks for the help in sorting this out

National Heritage list can be sourced to http://www.environment.gov.au/heritage/places/national-heritage-list
Commonwealth list can be sourced to http://www.environment.gov.au/heritage/places/commonwealth-heritage-list

the data I got from using the queries available at http://www.environment.gov.au/about-us/environmental-information-data/databases-applications extracting that data then filtering to only the necessary information, removing copyright material from the dataset as raw data(anything that can be sourced multiple ways) cant be copyrighted in Australia.

In T153221#3552463, @Gnangarra wrote:

the data I got from using the queries available at http://www.environment.gov.au/about-us/environmental-information-data/databases-applications extracting that data then filtering to only the necessary information, removing copyright material from the dataset as raw data(anything that can be sourced multiple ways) cant be copyrighted in Australia.

Could you provide an approximate date for this and maybe some deeper links for the queries? Pointing the url to either of the links in T153221#3552463 doesn't give users much help if they want to verify the info.

In T153221#3552183, @Effeietsanders wrote:

• For the identifier, I would suggest one single identifier, based on the 'entry in the Australian Heritage Database'. Possibly on top of that, we can create templates for the National Heritage List, Commonwealth Heritage List and register of the National Estate or even places under consideration. This would allow a single template for uploading, which can then automatically be replaced with the more specific template.

For the template :c:Template:Cultural Heritage Australia I would say start with a simple template taking only the place_id as the only parameter. It's fairly easy to add parameters to make the text/category change based on type afterwards and since the data will be in Wikidata one could write a bot to update anything which has already been uploaded.

In T153221#3552183, @Effeietsanders wrote:

(some kudos for Romaines supertable: https://commons.wikimedia.org/wiki/User:Romaine/Wiki_Loves_Monuments/2017/table )

• Category on Commons: seems to exist on https://commons.wikimedia.org/wiki/Category:Heritage_buildings_of_Australia, but this should get subtrees for the specific types of heritage, provinces, etc. This will need specifically two subcategories: one for the National Heritage List, and one for the Commonwealth Heritage List. [...]

I'm not sure that all items on the list are buildings. We might actually want a new category higher up in the Australian Cultural Heritage tree.

In T153221#3552409, @Gnangarra wrote:

Item 105694 is Lord Howe Island the first instance(424) is the area of the main island (Q104784) and the second instance(425) appears to be the area of the additional islands in the group which is covered by (Q1869866) and includes Wolf Rock(Q2695142), Sail Rock(no id) and Mutton Bird Island(no id) - combine as one entry under (Q1869866) which is the whole group of islands

Item 105698 - Macquarie Island (Q46650) - same as before multiple Islands though in this case there is no group, thought the first instance appears to be related to Judge and Clerk Islets (Q46439) & Bishop and Clerk Islets (Q46489) both which are part of the Macquarie Island Nature reserve (no WD item)

Item 105707 - combine as Heard Island and McDonald Islands (Q131198), the first instance is McDonald Island (Q1915097) and the second one is Heard Island (there is no WD id)

Thanks. I've ensured there is a group item for each and will simply not import area and coordinates for these automatically

In T153221#3552439, @Gnangarra wrote:

for 105891, 105704 and 106065

105891 - the Australian Alp NP, the different Coordinates are for individual Peaks within it
105704 - Gondwana Rain forest, different locations within the rain forest
106065 - two ships HMAS Sydney(Q1031986) @ 26°14′31″S 111°12′48″E, Kormoran(Q708046) @ 26°05′46″S 111°04′33″E both sank off Geraldton in WA following after attacking each other in WWII

Thanks. I'll simply drop automatically adding coordinates for these.

OK, created https://commons.wikimedia.org/wiki/Template:Cultural_Heritage_Australia and the related categories. As suggested by Lokal_Profil, I added heritage buildings in the category: https://commons.wikimedia.org/wiki/Category:Cultural_heritage_monuments_in_Australia

Have largely finished the code. You can see a preview of what the import would look like at https://www.wikidata.org/wiki/User:André_Costa_(WMSE)/COH/Australia/preview awaiting the final source components.

In T153221#3557167, @Lokal_Profil wrote:

In T153221#3552463, @Gnangarra wrote:

the data I got from using the queries available at http://www.environment.gov.au/about-us/environmental-information-data/databases-applications extracting that data then filtering to only the necessary information, removing copyright material from the dataset as raw data(anything that can be sourced multiple ways) cant be copyrighted in Australia.

Could you provide an approximate date for this and maybe some deeper links for the queries? Pointing the url to either of the links in T153221#3552463 doesn't give users much help if they want to verify the info.

around about 21 July 2017 was the date I sourced the data for both lists, that would accurate within a couple days either way the start point URL for the Australian Heritage list is http://data.gov.au/dataset/2016-soe-her-aus-national-heritage according to original csv file the data was last updated 31 May 2017

the commonwealth list was created at the same time from http://data.gov.au/dataset/commonwealth-heritage-list

In T153221#3557382, @Gnangarra wrote:

around about 21 July 2017 was the date I sourced the data for both lists, that would accurate within a couple days either way the start point URL for the Australian Heritage list is http://data.gov.au/dataset/2016-soe-her-aus-national-heritage according to original csv file the data was last updated 31 May 2017

the commonwealth list was created at the same time from http://data.gov.au/dataset/commonwealth-heritage-list

Thanks. I've now updated the sources and the preview example.

I created T174324: Allow Australia to participate in WLM2017 to more clearly separate Wiki-Loves-Monuments and Connected-Open-Heritage needs. I've also created a few subtasks there. For WLM related discussions I recommend continuing there and feeling free to unsubscribe here.