Re: [Wikidata] [Spam] Re: No links, wrong data: Scotland's orphans need help
Maybe there is a case to separate import and verification here? There are many statements in Wikidata nowadays, but they get really trustworthy through references (other than imported from Wikipedia). But for external IDs, references are superfluous; they are their own reference, by definition. So how about marking IDs with a verified (or last verified on) qualifier? Much of such work could be done by bots; we could then filter the problematic ones out for manual verification. As we have no control over external lists, this would have to be re-checked ever so often; but, again bots to the rescue. On Wed, Jun 3, 2015 at 12:49 PM Andrew Gray andrew.g...@dunelm.org.uk wrote: This particular case is something of a known problem - we've encountered it with some of the other heritage-building identifier lists as well. Bridges often span a river which is the border for two jurisdictions (in this case, council areas). Each local area counts it as a historic building, and because the national lists are aggregated from local lists, it gets two entries in the main list, one as Fife and one as Edinburgh. A similar case in Wales is the Menai Suspension Bridge, which is 4049 from the Gwynedd register and 18572 from the Anglesey one (Wikidata, at Q581526, only lists one identifer). The lack of deduplication is probably intentional rather than a bug, and both entries are correct. Perhaps one way to handle this for Wikidata would be to, hmm, say something like if the item is some kind of a bridge, then allow two IDs in the constraints? I can't immediately think of any bridges which cross national borders *and* are a heritage building in both countries, but we'd see the same thing there, with it having identifiers from both sides. Andrew. On 2 June 2015 at 12:12, Markus Krötzsch mar...@semantic-mediawiki.org wrote: Another interesting type of Scottish historic orphans are those that are duplicates of items that do have site links. Even very prominent ones are duplicated, such as https://www.wikidata.org/wiki/Q17569486 (dup) https://www.wikidata.org/wiki/Q933000 (real item) Interestingly, they use different Scotland IDs, and it does indeed seem that Historic Scotland also contains duplicates: http://data.historic-scotland.gov.uk/pls/htmldb/f?p=2200:15:0BUILDING,HL:47778 http://data.historic-scotland.gov.uk/pls/htmldb/f?p=2200:15:0BUILDING,HL:49165 Overall, this seems to be an example of an ID that really should not be considered identity providing since there seems to be an many-to-many relationship between Wikidata and Historic Scottland. Orphans should receive additional ids from a better source if at all possible. With the great number of seemingly legit non-functional uses of the Scotland IDs, they cannot be used in practice to detect duplicates. Regards, Markus On 02.06.2015 13:01, Markus Krötzsch wrote: On 02.06.2015 11:30, Magnus Manske wrote: Update 2: For example, https://www.wikidata.org/wiki/Q17847522 and https://www.wikidata.org/wiki/Q17847537 have the same Scotland ID, but refer to different entities (church and churchyard, respectively). They were as two entities in the original dataset, sharing the same ID. Yes, I noticed such cases too. From the information Wikidata, it is not clear to me why this is sometimes done and sometimes not done. For example, these adjacent houses have the same Scotland ID but different items that each have their own coordinates (where did the coordinates come from?): https://www.wikidata.org/wiki/Q17576211 https://www.wikidata.org/wiki/Q17576182 https://www.wikidata.org/wiki/Q17576185 In many other cases, adjacent houses with the same ID are combined into one item: https://www.wikidata.org/wiki/Q17806587 (note, however, that the house addresses given in the ID and in the item label do not match, though they overlap on most of the houses.) Finally, there are also cases where there are different IDs and we have several items, but they have the same labels that merge the contents of the two IDs: https://www.wikidata.org/wiki/Q17810121 https://www.wikidata.org/wiki/Q17810137 It seems that the data was not taken from the Historic Sites database but from some different source that has its own coordinate data and a different (but seemingly arbitrary) approach to grouping sites. However, the coordinated give Historic Scotland as their reference -- I wonder if Historic Scotland might be changing frequently or exist in several versions. Regards, Markus On Tue, Jun 2, 2015 at 10:26 AM Magnus Manske magnusman...@googlemail.com mailto:magnusman...@googlemail.com wrote: Update: There appear to be quite a few items with duplicate Scotland IDs (not all of them may be erroneous!): http://wdq.wmflabs.org/stats?action=doublestringprop=709 On Tue, Jun 2, 2015 at 10:23 AM
Re: [Wikidata] [Spam] Re: No links, wrong data: Scotland's orphans need help
On 03.06.2015 13:57, Magnus Manske wrote: Maybe there is a case to separate import and verification here? There are many statements in Wikidata nowadays, but they get really trustworthy through references (other than imported from Wikipedia). But for external IDs, references are superfluous; they are their own reference, by definition. So how about marking IDs with a verified (or last verified on) qualifier? Much of such work could be done by bots; we could then filter the problematic ones out for manual verification. As we have no control over external lists, this would have to be re-checked ever so often; but, again bots to the rescue. Yes, I fully support this proposal. What do you think about making last verified on not a qualifier but (part of) the reference information? The reference could state where the bot has looked up the ID and give a time. This would be somewhat similar to what is now used in Freebase Ids, e.g., in https://www.wikidata.org/wiki/Q42. In general, it might be useful to have such a last verified on property that can be added to arbitrary references. There are many other uses for this. One common case would be that a user has changed the value without even being aware of the reference -- then one would be able to detect this automatically by comparing the last modification time with the last verified on date. Putting the last verified on into the references also makes it possible to have different dates for different references there. Regards, Markus ___ Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata
Re: [Wikidata] [Spam] Re: No links, wrong data: Scotland's orphans need help
Thanks, Andrew, for the clarification. This makes perfect sense. I don't see a problem with one bridge having two IDs in some external database. We already have this for other ID-like properties for other reasons. What is important though is that it still is a single bridge, and should therefore be one item. Your clarification is reassuring since it suggests that the problem is not overly common after all. Maybe one can just merge these cases manually. Once the (multiple) ids are found in the merged items, avoiding future duplicates will be done as usual (which is still difficult with the Scottish Heritage ids since we have many legit Wikidata items that have the same id -- but this at least is an independent problem). Regards, Markus On 03.06.2015 13:48, Andrew Gray wrote: This particular case is something of a known problem - we've encountered it with some of the other heritage-building identifier lists as well. Bridges often span a river which is the border for two jurisdictions (in this case, council areas). Each local area counts it as a historic building, and because the national lists are aggregated from local lists, it gets two entries in the main list, one as Fife and one as Edinburgh. A similar case in Wales is the Menai Suspension Bridge, which is 4049 from the Gwynedd register and 18572 from the Anglesey one (Wikidata, at Q581526, only lists one identifer). The lack of deduplication is probably intentional rather than a bug, and both entries are correct. Perhaps one way to handle this for Wikidata would be to, hmm, say something like if the item is some kind of a bridge, then allow two IDs in the constraints? I can't immediately think of any bridges which cross national borders *and* are a heritage building in both countries, but we'd see the same thing there, with it having identifiers from both sides. Andrew. On 2 June 2015 at 12:12, Markus Krötzsch mar...@semantic-mediawiki.org wrote: Another interesting type of Scottish historic orphans are those that are duplicates of items that do have site links. Even very prominent ones are duplicated, such as https://www.wikidata.org/wiki/Q17569486 (dup) https://www.wikidata.org/wiki/Q933000 (real item) Interestingly, they use different Scotland IDs, and it does indeed seem that Historic Scotland also contains duplicates: http://data.historic-scotland.gov.uk/pls/htmldb/f?p=2200:15:0BUILDING,HL:47778 http://data.historic-scotland.gov.uk/pls/htmldb/f?p=2200:15:0BUILDING,HL:49165 Overall, this seems to be an example of an ID that really should not be considered identity providing since there seems to be an many-to-many relationship between Wikidata and Historic Scottland. Orphans should receive additional ids from a better source if at all possible. With the great number of seemingly legit non-functional uses of the Scotland IDs, they cannot be used in practice to detect duplicates. Regards, Markus On 02.06.2015 13:01, Markus Krötzsch wrote: On 02.06.2015 11:30, Magnus Manske wrote: Update 2: For example, https://www.wikidata.org/wiki/Q17847522 and https://www.wikidata.org/wiki/Q17847537 have the same Scotland ID, but refer to different entities (church and churchyard, respectively). They were as two entities in the original dataset, sharing the same ID. Yes, I noticed such cases too. From the information Wikidata, it is not clear to me why this is sometimes done and sometimes not done. For example, these adjacent houses have the same Scotland ID but different items that each have their own coordinates (where did the coordinates come from?): https://www.wikidata.org/wiki/Q17576211 https://www.wikidata.org/wiki/Q17576182 https://www.wikidata.org/wiki/Q17576185 In many other cases, adjacent houses with the same ID are combined into one item: https://www.wikidata.org/wiki/Q17806587 (note, however, that the house addresses given in the ID and in the item label do not match, though they overlap on most of the houses.) Finally, there are also cases where there are different IDs and we have several items, but they have the same labels that merge the contents of the two IDs: https://www.wikidata.org/wiki/Q17810121 https://www.wikidata.org/wiki/Q17810137 It seems that the data was not taken from the Historic Sites database but from some different source that has its own coordinate data and a different (but seemingly arbitrary) approach to grouping sites. However, the coordinated give Historic Scotland as their reference -- I wonder if Historic Scotland might be changing frequently or exist in several versions. Regards, Markus On Tue, Jun 2, 2015 at 10:26 AM Magnus Manske magnusman...@googlemail.com mailto:magnusman...@googlemail.com wrote: Update: There appear to be quite a few items with duplicate Scotland IDs (not all of them may be erroneous!): http://wdq.wmflabs.org/stats?action=doublestringprop=709 On Tue, Jun 2, 2015 at 10:23 AM Magnus Manske
