Wikidata:Property proposal/uses dataset
|Represents||data integration (Q386824), data analysis (Q1988917)|
|Template parameter||rarely or never in infobox|
|Domain||scholarly article (Q13442814)|
|Allowed values||only Wikidata items|
|Allowed units||no units|
|Example 1||Characteristics of oncology clinical trials: insights from a systematic analysis of ClinicalTrials.gov. (Q42600467) → AACT Database (Q76654384)|
|Example 2||Wiki Loves Monuments (Q1353202) → National Register of Historic Places (Q3719)|
|Example 3||New York Stock Exchange (Q13677) → (whatever the finance data is called?)|
|Example 4||Wikipedia (Q52) → Wikidata (Q2013)|
|Planned use||no immediate plans|
Can someone give me feedback on whether this is already proposed or if anyone has done something similar? I might pull this back and develop it more but I would like some feedback on how much this makes sense. Blue Rasberry (talk) 17:50, 4 February 2020 (UTC)
Historically researchers never shared their data. They would collect a dataset and publish their interpretation of results. Previously, publishing a dataset was difficult or challenge because paper media and even digital media used to be hard to share. A consequence of this was that research required trust that one person or team would do both data collection and interpretation, when these are often different skill sets. What is new now is that in the context of the Open Science Movement, researchers are deconstructing the research process. Now one researcher can present a dataset and another research can interpret it.
Research publication has a history of citing other papers. There are not widely-accepted standards for citing or crediting the use of databases. One common way to note the use of a dataset in a research project is to mention it in prose text of the body of a paper, and another way is to somehow make a nonstandard citation to a database after the model of citing an academic paper.
- Other uses
Lots of popular services use some datasets. Wikipedia does not currently have many articles on datasets, and Wikidata does not currently make many items for them, but I think this is worth exploring.
- weather websites, apps, and reports -> (some governmental source of weather data)
- politics and related journalism -> census data, demographic data, government datasets
- big websites -> Google search corpus, Amazon marketplace corpus, Facebook user profile corpus (none of these have names, I think)
- Comment Well we do have uses (P2283). Maybe some other existing properties like that could also work? ArthurPSmith (talk) 18:31, 5 February 2020 (UTC)
- Oppose I think uses (P2283) is sufficient. --Tinker Bell ★ ♥ 20:48, 6 February 2020 (UTC)
- Oppose - This proposal feels too broad for me. For specific claims references, using stated in (P248) is adequate (and can be linked to items). For large items (like New York Stock Exchange (Q13677)), there can be so many datasets that i don't really see the point. And for specific items, like Characteristics of oncology clinical trials: insights from a systematic analysis of ClinicalTrials.gov. (Q42600467), uses (P2283) seems good enough. Husky (talk) 18:05, 10 February 2020 (UTC)