SJK
Items about humans
editPlease note that most items about humans should have human (Q5) as value of instance of (P31). The most stuff you insert there now should be a occupation (P106). Sjoerd de Bruin (talk) 18:13, 14 February 2016 (UTC)
P3443
editSince Victorian Heritage Database ID (P3443) was created, I have e little question. What is the therm we sould use for heritage designation (P1435) are these have specific name like provincial historic resource (Q14465918) for Alberta Register of Historic Places ID (P759), or the are simply listed like Queensland Heritage Register ID (P2967)? --Fralambert (talk) 22:40, 7 January 2017 (UTC)
- @Fralambert: So, I created items to use as the "heritage status" value for the Victorian Heritage Database. All as subclasses of Victorian heritage status (Q28147628):
- listed on the Victorian Heritage Register (Q28147634) – for items listed on the Victorian Heritage Register (Q2221671) established by the Heritage Act 1995
- Victorian Heritage Inventory listing (Q28147635) – for items listed on the Victorian Heritage Inventory
- Heritage Overlay (Q17006517) – for heritage overlays established by local planning authorities
- National Trust of Australia (Victoria) listing (Q28147636) – for listings by the National Trust of Australia (Victoria)
- Victorian War Heritage Inventory listing (Q28147637)
- Note that some of these are statutory statuses (e.g. (1) and (3) are established by different provisions of state law), whereas (4) is not a legally binding status, it is maintained by a non-government body, but the government has elected to republish that information. (The legal status of (2) and (5) is unclear to me.)
- If you look at their database, on each item, on the right side of the screen, there is an expanadable "Record Details" section, if you expand that, look at "Heritage Listing" and "Level of Significance" fields, they give the heritage status of the item. Also note that the same item can be classified with multiple heritage statuses, in which case it gets separate entries in the DB. See for example ID# 123675 which is a National Trust listing (so not legally binding), yet the same place also has a Victorian Heritage Register listing under ID# 112535. SJK (talk) 00:48, 8 January 2017 (UTC)
Item to be delete
editIn RFD there are a lot of item proposed for the deletion created by you. If you do not agree you can participate in the debate --ValterVB (talk) 20:05, 3 March 2017 (UTC)
Harvesttemplate
editHi SJK,
You might want to try HarvestTemplate in Demo mode and then export the data to a spreadsheet and convert it to items there.
--- Jura 13:09, 4 March 2017 (UTC)
- Thanks but I came up with another way. I used Petscan to get the English wikipedia - drug article mapping into a file, then I wrote a shell script to download each article with curl (to get Special:Export), then I use grep/awk/sed/etc to extract pregnancy_category_AU= and pregnancy_category_US= from the article and turn it into QuickStatements. I'm running it all now.... SJK (talk) 13:24, 4 March 2017 (UTC)
- Yeah, after writing the above, I noticed you had started uploading them. Looks good.
--- Jura 13:56, 4 March 2017 (UTC)
- Yeah, after writing the above, I noticed you had started uploading them. Looks good.
Victorian state suburbs vs parishes
editHi SJK, I'm working on populating Australian Statistical Geography 2011 ID (P4014) which will help link our items to the 2011 census data about them. One batch of these codes are the "SSCs" which are state suburbs, and there is a Mix-N-Match set for them here. In Victoria I'm having a bit of trouble, as most of the places have two items with the same name, which is confusing the automatcher (and me). For example, for SSC21068 I was offered Paraparap (Q28931960), but ended up allocating it to Paraparap (Q18166285). Did I make the right choice? Should these items actually be merged, or is there a real difference? The ASGS also has another set of items which we haven't allocated yet, Gazetted Localities, here shown in a red box at GL_VIC2031. Which Wikidata item should that id be allocated to? (@Canley: you may also be interested in this discussion.) --99of9 (talk) 03:42, 21 June 2017 (UTC)
- @99of9:: I have been allocating mainly urban suburbs in mix-n-match as the correlation is usually good there, and there has been much discussion on the inaccuracy of SSCs in rural areas. Boundary-wise, they are very often not accurate, and the Gazetted Locality should be linked to the item. However, the ABS does not calculate population figures for GLs that don't correspond to an SSC, so that's what we have to work out – whether to list the SSC population with a note about all the GLs inside it, or whether it is close enough (population-wise) to link the named item to the SSC. As far as I know regarding the parishes, they are essentially all around the same size in Victoria and just named after a locality, so they won't correspond to the gazetted boundaries at all. I don't think the ABS has any corresponding structure, so I wouldn't link to these at all. In the example of Paraparap, the Paraparap GL follows the actual boundaries; the Paraparap SSC includes the area to the north around Moriac as well, but not including the actual Moriac UCL/SSC, and the Paraparap parish is a much larger square shape but with Thompson Creek as the northern boundary. --Canley (talk) 04:46, 21 June 2017 (UTC)
- @Canley:Ok I'm happy to ignore the parishes. Regarding SSC vs GL, for the moment I think both (and any other structure, e.g. UCL) should be linked within the Wikidata item. We can then decide to promote/deprecate/choose-data-from based on some criteria wherever we have multiple id's for the same item. --99of9 (talk) 05:05, 21 June 2017 (UTC)
- I agree, getting them matched in Wikidata is the main thing, once that's done it will be easier to assess quality and accuracy. --Canley (talk) 07:04, 21 June 2017 (UTC)
- @99of9: Canley has already covered much of this, and you may have worked it out yourself, but anyway: Parishes and localities/suburbs are different. Parishes are only used for recording the ownership of real estate (land titles). Localities/suburbs are used for numerous purposes. Localities/suburbs are the everyday concept and parishes are an obscurity only relevant to conveyancers and lovers of trivia. Even though they have the same name and are located in roughly the same area, their exact boundaries don't line up. (They probably used to line up more, but there can be everyday practical reasons like population growth/decline and cultural changes for altering names/boundaries of suburbs/localities, whereas no one wants to complicate land ownership records by changing the names/boundaries of the parishes.) I don't think it is proper to merge them because they are strictly speaking different things. I understand the potential for confusion (given that most people don't realise parishes exist), but I'm not sure what the solution is other than education. We could put the word "parish of" in the label, and that might make matching easier, but I think that violates Wikidata's normal conventions of using the descriptor for disambiguation instead. (Another solution is maybe if mix'n'match were enhanced to make the difference between two things have same label clearer.) I don't think linking parishes to statistical identifiers makes much sense because parishes are not used for most statistical purposes (statistics on land ownership might be an exception.)SJK (talk) 09:11, 22 June 2017 (UTC)
- Thanks, that's all very helpful. --99of9 (talk) 09:57, 22 June 2017 (UTC)
- @99of9: Canley has already covered much of this, and you may have worked it out yourself, but anyway: Parishes and localities/suburbs are different. Parishes are only used for recording the ownership of real estate (land titles). Localities/suburbs are used for numerous purposes. Localities/suburbs are the everyday concept and parishes are an obscurity only relevant to conveyancers and lovers of trivia. Even though they have the same name and are located in roughly the same area, their exact boundaries don't line up. (They probably used to line up more, but there can be everyday practical reasons like population growth/decline and cultural changes for altering names/boundaries of suburbs/localities, whereas no one wants to complicate land ownership records by changing the names/boundaries of the parishes.) I don't think it is proper to merge them because they are strictly speaking different things. I understand the potential for confusion (given that most people don't realise parishes exist), but I'm not sure what the solution is other than education. We could put the word "parish of" in the label, and that might make matching easier, but I think that violates Wikidata's normal conventions of using the descriptor for disambiguation instead. (Another solution is maybe if mix'n'match were enhanced to make the difference between two things have same label clearer.) I don't think linking parishes to statistical identifiers makes much sense because parishes are not used for most statistical purposes (statistics on land ownership might be an exception.)SJK (talk) 09:11, 22 June 2017 (UTC)
- I agree, getting them matched in Wikidata is the main thing, once that's done it will be easier to assess quality and accuracy. --Canley (talk) 07:04, 21 June 2017 (UTC)
- @Canley:Ok I'm happy to ignore the parishes. Regarding SSC vs GL, for the moment I think both (and any other structure, e.g. UCL) should be linked within the Wikidata item. We can then decide to promote/deprecate/choose-data-from based on some criteria wherever we have multiple id's for the same item. --99of9 (talk) 05:05, 21 June 2017 (UTC)