Hi Rick,

Well, I think you have a more financial-economic-investment focus on this topic than me. I have started the WikiObjects project proposal from a more costumer/user view,.

Anyway what you comment is quite interesting, I answer between lines:

2017-01-23 23:05 GMT+01:00 Rick Labs tmp2004@clbcm.com:

Hi Quico,

Thanks for your info about JOINS between PRODUCTS and COMPANIES. It gets me thinking of just how complex it can be.

Concerning PRODUCTS what instantly pops into my mind (with an accounting/finance background) is that the PRODUCT object should ideally fall singularly into one of the COMPANY's reported SEGMENTS (finest granular detail in public financial statements). Then there should be a reasonable ability to sum up sales by company by INDUSTRY. Then each industry should report singularly into a SECTOR, and the sum of all the sectors should agree with GROSS DOMESTIC PRODUCT.

Are you talking about the possibility of using wikidata for calculating GDP or similiar values? for example inclunding sales revenue values in each wikidata product item? If so, I think there are several previous stages to do before even planning this, like what you comment later: well-tagged granular data.

For example, here is one listing of INDUSTRIES: Aerospace & Defense Electronics | Airfreight & Surface Transportation | Airlines | Alternative Energy | Autos & Auto Parts | Business, Education & Professional Services | Chemicals | Electric Utilities | Electrical Equipment & Multi-Industry | Engineering & Construction | Food Producers | Food Retailers | Gaming & Lodging | Homebuilders & Building Products | Integrated Oil | Leisure | Machinery | Master Limited Partnerships (energy sector) | Metals & Mining | Paper & Packaging | Restaurants | Retailing/Broadlines & Hardlines | Retailing/Department Stores & Specialty Softlines | Semiconductors | Shipping | Software | Telecom & Networking Equipment.

The above need to be aggregated. For example the following economic SECTORS: Basic Materials | Capital Goods/Industrials | Consumer | Energy | Technology, Media & Telecommunications

There are some useful properties for tagging companies in wikidata, like "Industry" https://www.wikidata.org/wiki/Property:P452 and ISIC code https://www.wikidata.org/wiki/Property:P1796

International flow of goods (via product category, customs designations, import/export duties, etc.), regional, national and full global INPUT - OUTPUT tables, and analysis of transportation (freight at the regional, state, national and international levels) -- all very much rely on high quality PRODUCT classifications at the granular level.

For tagging products there are "Brand" https://www.wikidata.org/wiki/Property:P1716 , "by product of" https://www.wikidata.org/wiki/Property:P2822, or "production statatistics" P2746 https://www.wikidata.org/wiki/Property:P2746.

For product properties write a list in https://www.wikidata.org/wiki/Wikidata:WikiProject_Objects

Maybe there is some way for products to inherit properties from company , I mean, if Boing is a company from Airlines Industry , Boing 747 is a product from airlines industry.

The devil is in the details, especially when there is inter divisional sales, and inter company ownership. Accountants want to tabulate direct and indirect COSTS, leading to PROFIT/LOSS. Economists want to tabulate output and VALUE ADDED, for the most part unconcerned with INVENTORY buildup. The two models are difficult to fully reconcile without well tagged granular data.

Maybe first steps should go into ontology direction.

Another random thought is relating TRADE NAMES, BRANDS and PRODUCTS; along with generic and industry descriptions. PATENT by country is also important.

Properties for brand is "P1716 https://www.wikidata.org/wiki/Property:P1716", patent" is P1246 https://www.wikidata.org/wiki/Property:P1246 , for trade names should be necesary to dive into https://tools.wmflabs.org/hay/propbrowse/ searching for also for codenames , IDs , etc...

Then from some (distant in my case) exposure to manufacturing there are things like the RAW MATERIALS, BILL OF MATERIALS, EXPLOSION of the BoM into manufacturing steps (and lead times, labor requirement by JOB TITLE), ASSEMBLIES or MODULES, WORK IN PROCESS INVENTORY, and FINAL PRODUCTS.

BOM can achive with "Has part" property P527 https://www.wikidata.org/wiki/Property:P527 , here an exmaple of use https://tools.wmflabs.org/reasonator/?&q=27000417 https://tools.wmflabs.org/reasonator/?&q=27000417

I think of the practical aspects of auto parts - all those books and diagrams, plus all the vendors offering the same part - yet you still often need an experienced parts counter person (Sherpa / expert) to get the correct part on the first try. Wiper blades/spark plugs/oil filter/head lamps cross reference charts...over time a real nightmare.

Here is one of the places where I see a good application , like https://octopart.com/ but open source

Appliance manufacture and spare parts wholesaler "games" with part numbers to obfuscate reality and maintain proprietary sales flows.

Non OEM repair parts. Of the pillars of linked-data:

• People

• Places (geo-names)

• Organizations (corporate, public, non profit, trade, social....)

• Tme/data(Events)

• Products

I think Products is probably the very most difficult for free/open source to really address well. And then there are the services...

Hehe, incredible I also was thinking in services , but this is even more far , so step by step.

greetings

Rick

On 1/20/2017 6:46 AM, Quico Prol wrote:

I am interesed in one of its linkages, Products by a company (your last section) , I am started a wikiproject proposal about this https://meta.wikimedia.org/wiki/WikiObject there are several wikidata properties than joins products and companies like P859 https://www.wikidata.org/wiki/Property:P859 (Sponsor) and P176 https://www.wikidata.org/wiki/Property:P176 (Manufacturer) , It's an starting point.

2017-01-19 6:10 GMT+01:00 Rick Labs tmp2004@clbcm.com:

...

I'm taking a look at* Company Data* on WikiData and its use at WikiPedia. Interested in learning and/or formulating both the short term and long term plans.

I've created a very tentative and preliminary outline of the possible scope under my WikiData user space: https://www.wikidata.org/wiki/ User:Rjlabs/WikiData_Company_Data_Project

Looking for anyone interested in Company Data and its linkages. I have a good background in Company Data, and its analysis from a U.S. prospective but would love participation from UK, EU, Asian and other country company data experts. I especially need guidance from the WikiData developers/oncologists/project management regarding current and potential scope, and potential help with some of the detailed technical aspects.

Feel free to discuss here, email me, or post to the Discussion page related to the above link.

Suggestions as to how to kick this off and connect with all the correct people up front very much appreciated.

Thanks!

Rick Labs, CPA, CFA

---

Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata

---

Wikidata mailing listWikidata@lists.wikimedia.orghttps://lists.wikimedia.org/mailman/listinfo/wikidata

---

Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata

Eric,

Been pondering your question over the weekend and want to give a starter use-case as an example

I authored an article on WikiPedia: https://en.wikipedia.org/wiki/List_of_systemically_important_banks. Take a look at that an imagine an additional column "Total Assets". (There is more discussion of this on it's talk page).

See also:

The German version: https://de.wikipedia.org/wiki/Liste_global_systemrelevanter_Banken

The French version: https://fr.wikipedia.org/wiki/Banque_syst%C3%A9mique

Imagine, if they weren't sleeping here, Japanese, Chinese and other versions of the same

I envision all COMPANY DATA from those banks would actually be stored ONLY ONCE on WikiData. Then the chart would be exactly the same on the English, French and German version of the article on WikiPedia - only the chart LABELS would change to be expressed in English|French|German.

Use-case #1 How to create a table in WikiPedia that draws its COMPANY DATA (during the HTML page assembly) from WikiData? Also, how any user (French|German|English) could do the "data entry" task on ANY of those companies, and have that data only stored ONCE on WikiData?

Use-case #2 Same as above but would like to not only draw data from WikiData (such as total assets) but be able to store the total assets for each bank ONLY in it's native reported currency. Would like to TRANSFORM that SINGLE, "GOLD COPY" of the total assets data as stored in WikiData "ON THE FLY" (during HTML page assembly) into the native currency of the WikiPedia user using TODAYS EXCHANGE RATE. That exchange rate would also be stored, ONLY ONCE, in WikiData. The "transformation" here is easy, a simple multiplication.

There are many other technical challenges to be address, such as aligning ontologies across countries, dealing with conflicting external standards, linking challenges, basic datatypes for financial & economic data including time series and tic by tic data, dealing with dirty data, etc. Above is just an example, but very much on my mind as a challenge representative of a large class of problems in getting "Company Data" off on a good track at WikiData to facilitate moving forward.

Rick

On 1/20/2017 12:01 PM, Eric Scott wrote:

What technical aspects do you think you'll need help with?

On 01/18/2017 09:10 PM, Rick Labs wrote:

...

I especially need guidance from the WikiData developers/ontologists/project management regarding current and potential scope, and potential help with some of the detailed technical aspects.

---

Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata