Wikidata:Property proposal/Data interval

data interval

Originally proposed at Wikidata:Property proposal/Natural science

   Done: reports periodicity (P6339) (Talk and documentation)

Description	physical quantity corresponding to the interval between two measurements or data
Represents	time interval (Q186081) / periodicity of measurements (Q59635271)
Data type	Item
Domain	weather station (Q190107) automatic weather station (Q846837) weather history (P4150)
Allowed values	hourly measurement (Q59657010) daily measurement (Q59657036) monthly measurement (Q59657037) periodic climatology data (Q59657033)
Example 1	Bagotville A (Q42393053) → hourly measurement (Q59657010) (start time (P580) 1953)
Example 2	Bagotville A (Q42393053)/weather history (P4150) → monthly measurement (Q59657037)(start time (P580) 1953)
Example 3	Active Pass (Q59495897) → daily measurement (Q59657036) (start time (P580) 1984, end time (P582) 1996)
Source	Glossary - Climate - Environment and Climate Change Canada (see data interval and climate)
Planned use	Identify all 8750 stations of Meteorological Service of Canada (Q349450) with one or more values Qualify all the tab files for weather observations on Wiki Commons linked to every station using weather history (P4150) (monthly: 7668 files, daily: > 200K, hourly: > 400K, climate: ~8000)
Expected completeness	always incomplete (Q21873886)
See also	Meteorological Service of Canada climate site ID

Motivation

Weather stations, automatic and manual, are different one from another. Some are doing hourly observations, some daily, other are doing both. Some type of observations can be done during a short period of time, as reflected above in example 3. We need to be able to identify which station is doing which type of observations if one wants to use the observed value in a given context (ex: find all weather stations within 100 km radius of Montréal that had performed daily observation during the 80). This could not be done without this property.

Moreover, observations tab files located on Wiki Commons need to be qualified as well. Work will be done with WC developers to address this with categories and/or string within file path, but WD should be able to use properties to retrieve and identify files using common tools (instead of relying on string parsing through Lua, for example).

Dirac (talk) 17:55, 11 December 2018 (UTC)

Discussion

•  Comment A related property is event interval (P2257) - do you think this could be used for this purpose also without creating a new property? At the least it seems like we might want something a bit more general for this. ArthurPSmith (talk) 19:45, 13 December 2018 (UTC)

The idea is there. However, I don't feel that a weather observation could be qualified as an event. I am sensitive to the fact that the application could be more general though. My first draft was named data interval. It would applied to all type of data (ex: national survey, satellite data, etc.), and not specifically to weather observations. Do you think it would be more suited for a new property? Dirac (talk) 19:52, 13 December 2018 (UTC)

• I would  Support relabeling this proposal with the more general "data interval" concept, assuming that no existing property is suitable. ArthurPSmith (talk) 18:21, 14 December 2018 (UTC)

•  Comment there is also "publication interval", but that isn't the same either. --- Jura 09:30, 15 December 2018 (UTC)

•  Comment Based on these comments, I've updated the proposed property to correspond to time interval (Q186081), but generalized to any physical quantity, not only time. Another example could be the resolution of a satellite data set (distance). This would allow this property to characterize any data set. @Cantons-de-l'Est, Simon Villeneuve, Pigsonthewing, ArthurPSmith: Would this be a better match? Dirac (talk) 16:58, 18 December 2018 (UTC)

@Dirac, ArthurPSmith, Jura1:  Done: reports periodicity (P6339). − Pintoch (talk) 19:47, 6 January 2019 (UTC)

• Note that I don't see why this would be restricted to characterize weather stations, as this can apply to any climatic data, or geologic data, or astronomic data, or natural nd even technical data (series of measurements at regular intervals); for example a tachymeter can take mesurements at regular interval: the interval indicates the precision of time between measurements of observations. The mesurements themselves could be direct observations, or derived (e.g. averaged or cumulated over their characteristic interval); this could also apply to geopolitical data (e.g. census data). Do not confuse the interval of time (e.g. between two dates) during which measurements may be taken, with the periodicity of these measurements (observed directly or derived from observations by some average or cumulation), and note that the periodicity of mesasurements can change over history (e.g. yearly before 1950, monthly before 1995, daily before 2010, hourly after...), with also some periods during which measurements are not regular/periodic (but could be still be reported as estimated daily averages valid for all that period), and with the base of derived estimations (e.g. rain per dain, whose measurement unit could be mm/day or mm/year, independantly of the periods used report the data valid for a different period of time). So there are multiple "periods" or "time intervals" involved. We need somthing better. Some intervals of times will be specific (for example starting data ane ending date of operation of some data-collecting device or service). Also do not confuse these periods/intervals with the actual measurement that is not necessarily a measurement of time (could be measurement of wind speeds, rains, temperatures: the units for these estimated or cumulated measurement are very different). Verdy p (talk) 14:36, 4 June 2020 (UTC)