Wikidata:Property proposal/as percentage
Jump to navigation
Jump to search
as percentage[edit]
Not done
| Description | percentage of the parent value |
|---|---|
| Data type | Number (not available yet) |
| Domain | numeric value |
| Allowed values | +/-#,##0.0000 |
| Allowed units | percent |
| Example | senior (Q1358789) ÷ human population (Q33829) × 100 → xx.xx |
| See also | water as percent of area (P2927) |
- Motivation – Needful qualifier for vote and election results
- eligible voters (P1867) 1,745,635
- ballots cast (P1868) 1,244,987
- as percentage (Pxxxxx) 71.32
qualifier
- as percentage (Pxxxxx) 71.32
- total valid votes (P1697) 1,220,394
I found no solution saving a percentaged value related to a numeric value
- Discussion
New proposal --Plagiat (talk) 17:41, 24 June 2016 (UTC)
Comment this seems like the sort of thing that should be calculated as required rather than stored as a property? Thryduulf (talk) 00:44, 25 June 2016 (UTC)
- I thought about this also. But as example: I generate user defined voting reports from municipalities of Bavaria (Germany), the maximum is round about 6,400 calculations. I am not sure, what way is faster and cheaper, calculate or query. --Plagiat (talk) 15:54, 25 June 2016 (UTC)
- Comment Maybe I may use conversion to standard unit (P2442) --Plagiat (talk) 16:23, 25 June 2016 (UTC)
Info Additional argument: Querying a single electoral constituency delivers all needed / available data. --Plagiat (talk) 11:05, 26 June 2016 (UTC)
Oppose This can be calculated. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:14, 29 June 2016 (UTC)- Oppose I don't think wikidata should be used as a store to speed up presentation of data. Even a million calculations would be less than a second on just about any device these days. Maybe if there are properties where this is a common issue we could have a javascript gadget that does the calculation as a convenience for people looking at wikidata. Anyway, no I don't think this should be a property. ArthurPSmith (talk) 17:55, 30 June 2016 (UTC)
Support can't necessarily be calculated
--- Jura 17:57, 30 June 2016 (UTC)- @Jura1: In which cases can it not be calculated? --Yair rand (talk) 22:32, 5 July 2016 (UTC)
- Support If there are two population numbers from two different sources and two GDP numbers from two different sources it's not trival to calculate the percentage.
There also might be cases where the uncertainty of a percentage that a source provides is lower than the uncertainty that you get by strict calculation. ChristianKl (talk) 19:40, 8 July 2016 (UTC)- Info I have only examples at German WP CDU Bundestagswahlen and CSU Europawahlergebnisse --Plagiat (talk) 11:21, 7 July 2016 (UTC)
- Comment "as percentage" of what? "valid votes" out of "ballots cast", "eligible voters", "registered voters" or out of what? Yes, we probably need something like this, but this model does not look firm enough! -- Innocent bystander (talk) 18:46, 29 July 2016 (UTC)
- That could be solved by using relative to (P2210) as qualifier. Lymantria (talk) 17:17, 2 August 2016 (UTC)
- @Lymantria: I do not think that works. We do not allow qualifiers to qualifiers. Take 2016 United Kingdom European Union membership referendum (Q21812812) as an example. UK has a population of ~64M people. 46 500 001 had the right to vote. 33 577 342 used that right. 33 551 983 ballots were valid. 16 141 241 voted "stay", 17 410 742 voted "leave". That is 51,9 % of valid votes, 51,8 % of total votes, 37,4 % of those who had the right to vote and 27,2 % of the total population. To be useful, this new property have to be a qualifier itself. And relative to (P2210) have to be a qualifier to that qualifier. Otherwise wo do not know which "as percentage" belongs to which number. The only alternative is then to allow only one "as percentage" for each number. If you compare with more than one number, the data looses its integrity. -- Innocent bystander (talk) 19:08, 2 August 2016 (UTC)
- @Innocent bystander: You are right, I had been shortsighted here. I have the same concern you have. I thought this qualifier would solve that, but that isn't the case. A percentage has no meaning if it is not defined relative to something. Lymantria (talk) 19:39, 2 August 2016 (UTC)
- I would like to open up to one (or more) less generic properties here. "percentage of valid ballots" or "percentage of casted ballots". We already have water as percent of area (P2927) which is not very generic. We also have to consider those cases which cannot be "calculated". That candidate "A" got 15% is maybe the only thing we know in some cases. I have hundred of Swedish municipal elections mentioned on svwiki where we do not know how many voted, how many ballots each candidate got etc. All we know is how many seats each political party got. In many cases all the seats were taken by independent candidates. -- Innocent bystander (talk) 11:51, 3 August 2016 (UTC)
- We also have central government debt as a percent of GDP (P1689). Your idea is fine with me. Perhaps it is better not to refer to ballots, to allow percentages on votes when computers are/were used. Lymantria (talk) 21:19, 3 August 2016 (UTC)
- I would like to open up to one (or more) less generic properties here. "percentage of valid ballots" or "percentage of casted ballots". We already have water as percent of area (P2927) which is not very generic. We also have to consider those cases which cannot be "calculated". That candidate "A" got 15% is maybe the only thing we know in some cases. I have hundred of Swedish municipal elections mentioned on svwiki where we do not know how many voted, how many ballots each candidate got etc. All we know is how many seats each political party got. In many cases all the seats were taken by independent candidates. -- Innocent bystander (talk) 11:51, 3 August 2016 (UTC)
- @Innocent bystander: You are right, I had been shortsighted here. I have the same concern you have. I thought this qualifier would solve that, but that isn't the case. A percentage has no meaning if it is not defined relative to something. Lymantria (talk) 19:39, 2 August 2016 (UTC)
- @Lymantria: I do not think that works. We do not allow qualifiers to qualifiers. Take 2016 United Kingdom European Union membership referendum (Q21812812) as an example. UK has a population of ~64M people. 46 500 001 had the right to vote. 33 577 342 used that right. 33 551 983 ballots were valid. 16 141 241 voted "stay", 17 410 742 voted "leave". That is 51,9 % of valid votes, 51,8 % of total votes, 37,4 % of those who had the right to vote and 27,2 % of the total population. To be useful, this new property have to be a qualifier itself. And relative to (P2210) have to be a qualifier to that qualifier. Otherwise wo do not know which "as percentage" belongs to which number. The only alternative is then to allow only one "as percentage" for each number. If you compare with more than one number, the data looses its integrity. -- Innocent bystander (talk) 19:08, 2 August 2016 (UTC)
- I'm not sure if there is a universal way to determine these percentages. There may be slight differences on what is included or excluded. So if you do want to describe it, you may want an additional qualifier, even if it's parallel to this one.
--- Jura 06:31, 8 August 2016 (UTC)
- That could be solved by using relative to (P2210) as qualifier. Lymantria (talk) 17:17, 2 August 2016 (UTC)
@Plagiat, Thryduulf, Pigsonthewing, Jura1, Innocent bystander, Lymantria: 
Not done, no consensus. This should be calculated. When we find actual data that can not be calculated, this might convince people that this property is needed. --Srittau (talk) 11:58, 23 August 2016 (UTC)