Dear All,

it is great that we are having this discussion, but may I please suggest to have it on the RfP page on Wikidata? People already asked similar questions there, and, in my experience, on-wiki discussion will likely lead to refined request which will accomodate all suggestions.

Cheers
Yaroslav

On Mon, Oct 16, 2017 at 5:53 PM, Sebastian Hellmann <hellmann@informatik.uni-leipzig.de> wrote:

ah, ok, sorry, I was assuming that Blazegraph would transitively resolve this automatically.

Ok, so let's divide the problem:

# Task 1:

Connect all existing organisations with the data from the handelsregister. (No new identifiers added, we can start right now)

Add a constraint that all German organisations should be connected to a court, i.e. the registering organisation as well as the id assigned by the court.

@all: any properties I can reuse for this?

I will focus on this as it seems quite easy. We can first filter orgs by other criteria, i.e. country as a blocking key and then string match the rest.

# Task 2:

Add all missing identifiers for the remaining orgs in Handelsregister. Whereas 2 can be rediscussed and decided, if 1 is finished sufficiently.


# regarding maintenance:
I find Wikidata as such very hard to maintain as all data is copied from somewhere else eventually, but Wikipedia has the same problem. In the case of the German Business register, maintenance is especially easy as the orgs are stable and uniquely identifiable. Even the fact that a company gets shut down should still be in Wikidata, so you have historical information. I mean, you also keep the Roman Empire, the Hanse and even finished projects in Wikidata. So even if an org ceases to exist, the entry in Wikidata should stay.

# regarding Opencorporates
I have a critical opinion with Opencorporates. It appears to be open, but you actually can not get the data. If somebody has a data dump, please forward to me. Thanks.
More on top, I consider Opencorporates a danger to open data. It appears to push open availability of data, but then it is limited to open licenses. Usefulness is limited as there are no free dumps and no possibility to duplicate it effectlively. Wikipedia and Wikidata provide dumps and an API for exactly this reason. Everytime somebody wants to create an open organisation dataset with no barriers, the existence of Opencorporates is blocking this.

Cheers,
Sebastian


On 16.10.2017 15:34, Antonin Delpeuch (lists) wrote:
And… my own count was wrong too, because I forgot to add DISTINCT in my
query (if there are multiple paths from the class to "organization
(Q43229)", items will appear multiple times).

So, I get 1 168 084 now.
http://tinyurl.com/yaeqlsnl

It's easy to get these things wrong!

Antonin

On 16/10/2017 14:16, Antonin Delpeuch (lists) wrote:
Thanks Ettore for spotting that!

Wikidata types (P31) only make sense when you consider the "subclass of"
(P279) property that we use to build the ontology (except in a few cases
where the community has decided not to use any subclass for a particular
type).

So, to retrieve all items of a certain type in SPARQL, you need to use
something like this:

?item wdt:P31/wdt:P279* ?type

You can also have other variants to accept non-truthy statements.

Just with this truthy version, I currently get 1 208 227 items. But note
that there are still a lot of items where P31 is not provided, or
subclasses which have not been connected to "organization (Q43229)"…

So in general, it's very hard to have any "guarantees that there are no
duplicates", just because you don't have any guarantees that the
information currently in Wikidata is complete or correct.

I would recommend trying to import something a bit smaller to get
acquainted with how Wikidata works and what the matching process looks
like in practice. And beyond a one-off import, as Ettore said it is
important to think how the data will be maintained in the future…

Antonin

On 16/10/2017 13:46, Ettore RIZZA wrote:
    - Wikidata has 40k organisations: 

    https://query.wikidata.org/#SELECT
    <https://query.wikidata.org/#SELECT> %3Fitem %3FitemLabel %0AWHERE
    %0A{%0A %3Fitem wdt%3AP31 wd%3AQ43229.%0A SERVICE wikibase%3Alabel {
    bd%3AserviceParam wikibase%3Alanguage "[AUTO_LANGUAGE]%2Cen". }%0A}


Hi, 

I think Wikidata contains many more organizations than that. If we
choose the "instance of Business enterprise", we get 135570 results. And
I imagine there are many other categories that bring together commercial
companies.


https://query.wikidata.org/#SELECT%20%3Fitem%20%3FitemLabel%20WHERE%20%7B%0A%20%20%3Fitem%20wdt%3AP31%20wd%3AQ4830453.%0A%20%20SERVICE%20wikibase%3Alabel%20%7B%20bd%3AserviceParam%20wikibase%3Alanguage%20%22%5BAUTO_LANGUAGE%5D%2Cen%22.%20%7D%0A%7D

On the substance, the project to add all companies of a country would
make Wikidata a kind of totally free clone of Open Corporates
<https://opencorporates.com/>. I would of course be delighted to see
that, but is it not a challenge to maintain such a database? Companies
are like humans, it appears and disappears every day.

 

2017-10-16 13:41 GMT+02:00 Sebastian Hellmann
<hellmann@informatik.uni-leipzig.de
<mailto:hellmann@informatik.uni-leipzig.de>>:

    Hi all,

    the technical challenges are not so difficult.

    - 2.2 million are the exact number of German organisations, i.e.
    associations and companies. They are also unique.

    - Wikidata has 40k organisations:

    https://query.wikidata.org/#SELECT
    <https://query.wikidata.org/#SELECT> %3Fitem %3FitemLabel %0AWHERE
    %0A{%0A %3Fitem wdt%3AP31 wd%3AQ43229.%0A SERVICE wikibase%3Alabel {
    bd%3AserviceParam wikibase%3Alanguage "[AUTO_LANGUAGE]%2Cen". }%0A}

    so there would be a maximum of 40k duplicates These are easy to find
    and deduplicate

    - The crawl can be done easily, a colleague has done so before.  


    The issues here are:

    - Do you want to upload the data in Wikidata? It would be a real big
    extension. Can I go ahead

    - If the data were available externally as structured data under
    open license, I would probably not suggest loading it into wikidata,
    as the data can be retrieved from the official source directly,
    however, here this data will not be published in a decent format.

    I thought that the way data is copied from coyrighted sources, i.e.
    only facts is ok for wikidata. This done in a lot of places, I
    guess. Same for Wikipedia, i.e. News articles and copyrighted books
    are referenced. So Wikimedia or the Wikimedia community are experts
    on this.

    All the best,

    Sebastian


    On 16.10.2017 10:18, Neubert, Joachim wrote:
    Hi Sebastian,____

    __ __

    This is huge! It will cover almost all currently existing German
    companies. Many of these will have similar names, so preparing for
    disambiguation is a concern.____

    __ __

    A good way for such an approach would be proposing a property for
    an external identifier, loading the data into Mix-n-match,
    creating links for companies already in Wikidata, and adding the
    rest (or perhaps only parts of them - I’m not sure if having all
    of them in Wikidata makes sense, but that’s another discussion),
    preferably with location and/or sector of trade in the description
    field.____

    __ __

    I’ve tried to figure out what could be used as key for a external
    identifier property. However, it looks like the registry does not
    offer any (persistent) URL to its entries. So for looking up a
    company, apparently there are two options:____

    __ __

    -          conducting an extended search for the exact string “A&A
    Dienstleistungsgesellschaft mbH“____

    -          copying the register number “32853” plus selecting the
    court (Leipzig) from the according dropdown list and search that____

    __ __

    Both ways are not very intuitive, even if we can provide a link to
    the search form. This would make a weak connection to the source
    of information. Much more important, it makes disambiguation in
    Mix-n-match difficult. This applies for the preparation of your
    initial load (you would not want to create duplicates). But much
    more so for everybody else who wants to match his or her data
    later on. Being forced to search for entries manually in a
    cumbersome way for disambiguation of a new, possibly large and
    rich dataset is, in my eyes, not something we want to impose on
    future contributors. And often, the free information they find in
    the registry (formal name, register number, legal form, address)
    will not easily match with the information they have (common name,
    location, perhaps founding date, and most important sector of
    trade), so disambiguation may still be difficult.____

    __ __

    Have you checked which parts of the accessible information as
    below can be crawled and added legally to external databases such
    as Wikidata?____

    __ __

    Cheers, Joachim____

    __ __

    --____

    Joachim Neubert____

    __ __

    ZBW – German National Library of Economics____

    Leibniz Information Centre for Economics____

    Neuer Jungfernstieg 21
    20354 Hamburg____

    Phone +49-42834-462____

    __ __

    __ __

    __ __

    *Von:*Wikidata [mailto:wikidata-bounces@lists.wikimedia.org
    <mailto:wikidata-bounces@lists.wikimedia.org>] *Im Auftrag von
    *Sebastian Hellmann
    *Gesendet:* Sonntag, 15. Oktober 2017 09:45
    *An:* wikidata@lists.wikimedia.org
    <mailto:wikidata@lists.wikimedia.org>
    *Betreff:* [Wikidata] Kickstartet: Adding 2.2 million German
    organisations to Wikidata____

    __ __

    Hi all,____

    the German business registry contains roughly 2.2 million
    organisations. Some information is paid, but other is public, i.e.
    the info you are searching for at and clicking on UT (see example
    below):____

    https://www.handelsregister.de/rp_web/mask.do?Typ=e
    <https://www.handelsregister.de/rp_web/mask.do?Typ=e>____

    __ __

    I would like to add this to Wikidata, either by crawling or by
    raising money to use crowdsourcing concepts like crowdflour or
    amazon turk. ____

    __ __

    It should meet notability criteria 2:
    https://www.wikidata.org/wiki/Wikidata:Notability
    <https://www.wikidata.org/wiki/Wikidata:Notability>____

        2. It refers to an instance of a *clearly identifiable
        conceptual or material entity*. The entity must be notable, in
        the sense that it *can be described using serious and publicly
        available references*. If there is no item about you yet, you
        are probably not notable.____


    The reference is the official German business registry, which is
    serious and public. Orgs are also per definition clearly
    identifiable legal entities.

    How can I get clearance to proceed on this?

    All the best,
    Sebastian____

    __ __

    __ __


          Entity data____

    __ __

    Saxony District court *Leipzig HRB 32853 * – A&A
    Dienstleistungsgesellschaft mbH ____

    Legal status:____

    	

    Gesellschaft mit beschränkter Haftung  ____

    	
    	

    Capital:____

    	

    25.000,00 EUR ____

    	
    	

    Date of entry:____

    	

    29/08/2016
    (When entering date of entry, wrong data input can occur due to
    system failures!) ____

    	
    	

    Date of removal:____

    	

    - ____

    	
    	

    Balance sheet available: ____

    	

    - ____

    	
    	

    Address (subject to correction):____

    	

    A&A Dienstleistungsgesellschaft mbH
    Prager Straße 38-40____

    04317 Leipzig ____

    	
    	

    __ __

    -- 
    All the best,
    Sebastian Hellmann

    Director of Knowledge Integration and Linked Data Technologies
    (KILT) Competence Center
    at the Institute for Applied Informatics (InfAI) at Leipzig University
    Executive Director of the DBpedia Association
    Projects: http://dbpedia.org, http://nlp2rdf.org,
    http://linguistics.okfn.org, https://www.w3.org/community/ld4lt
    <http://www.w3.org/community/ld4lt>
    Homepage: http://aksw.org/SebastianHellmann
    <http://aksw.org/SebastianHellmann>
    Research Group: http://aksw.org____



    _______________________________________________
    Wikidata mailing list
    Wikidata@lists.wikimedia.org <mailto:Wikidata@lists.wikimedia.org>
    https://lists.wikimedia.org/mailman/listinfo/wikidata
    <https://lists.wikimedia.org/mailman/listinfo/wikidata>
    -- 
    All the best,
    Sebastian Hellmann

    Director of Knowledge Integration and Linked Data Technologies
    (KILT) Competence Center
    at the Institute for Applied Informatics (InfAI) at Leipzig University
    Executive Director of the DBpedia Association
    Projects: http://dbpedia.org, http://nlp2rdf.org,
    http://linguistics.okfn.org, https://www.w3.org/community/ld4lt
    <http://www.w3.org/community/ld4lt>
    Homepage: http://aksw.org/SebastianHellmann
    <http://aksw.org/SebastianHellmann>
    Research Group: http://aksw.org

    _______________________________________________
    Wikidata mailing list
    Wikidata@lists.wikimedia.org <mailto:Wikidata@lists.wikimedia.org>
    https://lists.wikimedia.org/mailman/listinfo/wikidata
    <https://lists.wikimedia.org/mailman/listinfo/wikidata>




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

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

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


--
All the best,
Sebastian Hellmann

Director of Knowledge Integration and Linked Data Technologies (KILT) Competence Center
at the Institute for Applied Informatics (InfAI) at Leipzig University
Executive Director of the DBpedia Association
Projects: http://dbpedia.org, http://nlp2rdf.org, http://linguistics.okfn.org, https://www.w3.org/community/ld4lt
Homepage: http://aksw.org/SebastianHellmann
Research Group: http://aksw.org

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