Wikidata talk:WikiProject Economics
We were just disussing this topic at de:Portal Diskussion:Wirtschaft when you left your note there, especially regarding the date of the information. For example, see de:Siemens: The employee count is of September 30, 2012 and the revenue is for the fiscal year 2011/12. Both data are only meaningful if one knows this date attribute. And they may not relate to the same date - e.g. we may have a current employee count of June 1 2014, but the revenue will refer to the year ending on December 31, 2013. Did you consider this date attribute in your concept? --PM3 (talk) 21:34, 23 June 2014 (UTC)
- Oh, I just notice that this project is not about company data but just about ... what's the English term for "Volkswirtschaft"? political economics? I am not into that, but I think that the basic problem is the same: It is important to know the date of information. --PM3 (talk) 21:39, 23 June 2014 (UTC)
Another issue is that in an encyclopedia, historical data have the same notability as current data. So it would be poor if we store only the last availabe information in Wikidata and discard the historical data. I would prefer very much if we can store not only one revenue information for each company, but multiple records for multiple fiscal years. Did you consider this? --PM3 (talk) 21:34, 23 June 2014 (UTC)
Many economic properties were created today: Wikidata:Property_proposal/Pending/2 (not by me). Can someone from this project go through the list and update the tables of this project? --Tobias1984 (talk) 14:59, 4 October 2015 (UTC)
- @Mcnabber091: --Tobias1984 (talk) 15:03, 4 October 2015 (UTC)
- @Tobias1984: Hi Tobias, sorry for my inactivity. I will update the tables.Mcnabber091 (talk) 19:15, 11 November 2015 (UTC)
- @Mcnabber091: No problem. Hopefully the introduction of quantities will draw more people into this project. And https://maps.wikimedia.org/ will hopefully soon give us the possibility to put the data on a map. --Tobias1984 (talk) 19:32, 11 November 2015 (UTC)
- @Tobias1984: Thanks Tobias, just so you know I have been busy working on http://econfactbook.org/data.html which is very related this project. I hope that Econ Factbook can be used as a model for how we should organize economic data on Wikidata. Please click on one of the countries to see the format. Mcnabber091 (talk) 01:35, 12 November 2015 (UTC)
- @Tobias1984: I just made corrections to the pending page. I made all the edits that I thought were need. Please let me know if more is needed Mcnabber091 (talk) 03:18, 12 November 2015 (UTC)
- @Mcnabber091: The page should ideally be empty after all properties are either archived or created. I currently would like to slowly progress with creating the economic properties. Some of the existin properties are rarely used and I want to avoid bad scaffolding of the economic properties. I think after a bot has imported a good chunk of data it will be easier to see what is missing and how best to do it. --Tobias1984 (talk) 21:37, 12 November 2015 (UTC)
It seems that in the meantime, there is a decent amount of properties, but these are hardly used. While I understand that the data quantity isn't easy to use nor fully implemented, it would be interesting to see a couple of sample items. --- Jura 13:12, 21 December 2015 (UTC)
System of National Accounts & Statistical Business RegistryEdit
Over at Wikidata:WikiProject_Companies we are in the early stages of looking at refactoring the class hierarchy starting with organization (Q43229) and going down. Problems and challenges abound. For example directly under organization (Q43229) there are a pig's breakfast of classes. This starts with classes like:
- economic entity (Q12569864)
- institution (Q178706)
- financial institution (Q650241)
- operation (Q362482)
- company (Q783794)
- enterprise (Q6881511)
- business (Q4830453)
- institute (Q1664720)
- organizational subdivision (Q9261468)
- front organization (Q1470315)
- non-governmental organization (Q79913)
- nonprofit organization (Q163740)
- association (Q15911314)
- international organization (Q484652)
- meritocracy (Q178079)
Contains duplication and overlap, such as:
- professional body (Q1110684)
- political action committee (Q2101438)
- pressure group (Q1666019)
Contains less economically significant entities in detail while missing huge sections of the economy (e.g. oil and gas, transportation, steel, utilities, cars, etc.):
Contains odd organizations but lacks all the basics. Run this query to see all classes under Organization User:Rjlabs/WikiData_Company_Data_Project#All_subclasses_of_a_class_example_1
A solution would be to refactor the classes such that a global Statistical Business Register (in the classic economic sense) could easily be extracted. One goal is to be able to extract all organizations that contribute to GDP and add them up without duplication. Another goal would be to be able to extract them by SNA categories. Help is needed (particularly economic experts, SNA experts, SBR experts) to get this right at an early stage. As more and more "company data" flows in to wikipeda, the harder it will be to make wholesale changes to its class structure without push back. We basically need to build a solid SNA compatible class structure for organizations. (Or...if you have a better, more global approach, by all means lets here it. Rjlabs (talk) 15:04, 1 April 2017 (UTC)
Currency for Properties like nominal GDP (P2131) etc.Edit
Properties like "nominal GDP" P2131 are usually available in US-Dollar and in the local currency - for Example euro for Countries in the Eurozone like France, Germany etc. The World-Bank offers the values for both currencies:
I propose to use both values for the property P2131 - USD and Local currency. In this case for almost all the countries there should be 2 entries for each year.
Alternatively we could use another property for the nominal GDP in the Local Currency.
Personally I prefer the existing property P2131 using the currency (for example USD, euro etc.) as unit.
WorldBank Data - Using the API for WikidataEdit
Notified participants of WikiProject Economics
The world bank offers a API to access data like GDP etc. Here you can find the description:
Here you can find an example for the nominal GDP in USD for the United States:
I propose a Wikidata-Bot with these properties:
- Read and Write new entries for all the Countries, connecting to the API
- Delete / Update Dada regularly
Foreign direct investment net inflow and outflow - specification neededEdit
The Properties Property:P2141 and Property:P2140 do not have any statements. I would like to add some using the bot User:WDBot. I was preparing to do test edits and I found, that the original description of the discussion about the proposal of the property Wikidata:Property proposal/Archive/37#P2141 does not correspond to the source of the data. The last sentence of the description of the proposal says [...] This series shows net outflows of investment from the reporting economy to the rest of the world and is divided by GDP. In source values https://data.worldbank.org/indicator/BX.KLT.DINV.CD.WD are absolute values in US-Dollar (Foreign direct investment, net inflows (BoP, current US$).
Which is here the relevant value (here net inflows)?
- https://data.worldbank.org/indicator/BX.KLT.DINV.CD.WD - Foreign direct investment, net inflows (BoP, current US$) or
- https://data.worldbank.org/indicator/BX.KLT.DINV.WD.GD.ZS - Foreign direct investment, net inflows (% of GDP)?
- @Datawiki30: The properties are Quantity datatype, so you could potentially use either approach and just specify the unit (as US$ or % of GDP). However, I think US$ or whatever the provided currency unit is in the data source should be the default for this kind of data. ArthurPSmith (talk) 13:26, 26 October 2018 (UTC)
Please see Wikidata:Bot_requests#Add_annual_country_level_unemployment_rate_(P1198). --- Jura 13:36, 26 October 2018 (UTC)
Discussion about the use and modeling of properties for statistical dataEdit
Hi, here you can find one discussion about the use of properties for statistical data: https://m.wikidata.org/wiki/Wikidata:Project_chat#Some_fundamental_questions_about_modeling_properties_for_statistical_data
- Dear all, after some discussion I have started a voting for consensus. Please see here. Cheers! --Datawiki30 (talk) 19:05, 4 November 2018 (UTC)
Hi everybody! Here I would like to make a try to summarize some main results from the discussions about the modeling of times-series data like nominal GDP (P2131):
- With the actual Wikidata UI it is not a good idea to import large time series like I have done this for the nominal GDP, see here
- There is now a Lua function that can access wikidata entries depending on the reference and unity!
- We have now a good Lua function that can scale for example big integers/numers. See here
- With the help of the ListeriaBot we can make list from wikidata queries. Some development in design etc. is still needed to replicate this for example. Here you can see one first "raw" result.
The most important issue at the moment is the performance of the Wikidata UI, that suffers when we have many statements like times-series. I think that data like pupulation, GDP etc. is important for Wikidata. Bacause of the performance issue I would propose to import only the most actual data for times-series-properties like population, gdp, unemployment, inflation - practically almost all the properties mentioned on the Economics project's page ... Any consensus about this?
CNET electronics specsEdit
The CNET website has structured specs for all sorts of electronics products, with an URL structure like https://www.cnet.com/products/lenovo-thinkpad-x200/specs/ for ThinkPad X200 (Q51954707). How many items for such products do we already have, that could use e.g. some automated verification of their content? They wayback machine doesn't have that many, slightly more than 60k URLs, but at least it's a start to see what there is. https://web.archive.org/web/*/https://www.cnet.com/products/* --Nemo 07:20, 30 October 2018 (UTC)