Re: [Wikidata] Photographers' Identities Catalog (& WikiData)

2015-12-15 Thread John Erling Blad 
There are some pretty good methods for optimizing the match process, but I
have not seen any implementation for that against Wikidata items. Only
things I've seen are some opportunistic methods. Duck tests gone wrong, or
"Darn it was a platypus!"

On Mon, Dec 14, 2015 at 11:19 PM, André Costa 
wrote:

> I'm planning to bring a few of the datasets into mix'n'match (@Magnus this
> is the one I asked sbout on Twitter) in January but not all of them are
> suitable and I believe separating KulturNav into multiple datasets on
> mix'n'match maxes more sense and makes it more likely that they get matched.
>
> Some of the early adopters of KulturNav have been working with WMSE to
> facilitate bi-directional matching. This is done on a dataset-by-dataset
> level since different institutions are responsible for different datasets.
> My hope is that mix'n'match will help in this area as well, even as a tool
> for the institutions own staff who are often interested in matching entries
> to Wikipedia (which most of the time means wikidata).
>
> @John: There are processes for matching kulturnav identifiers to wikidata
> entities. Only afterwards are details imported. Mainly to source statements
> [1] and [2]. There is some (not so user friendly) stats at [3].
>
> Cheers,
> André
>
> [1]
> https://www.wikidata.org/wiki/Wikidata:Requests_for_permissions/Bot/L_PBot_2
> [2]
> https://www.wikidata.org/wiki/Wikidata:Requests_for_permissions/Bot/L_PBot_3
> [3] https://tools.wmflabs.org/lp-tools/misc/data/
> --
> André Costa
> GLAM developer
> Wikimedia Sverige
>
> Magnus Manske, 13/12/2015 11:24:
>
> >
> > Since no one mentioned it, there is a tool to do the matching to WD much
> > more efficiently:
> > https://tools.wmflabs.org/mix-n-match/
> 
>
> +1
>
> ___
> 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

Re: [Wikidata] Photographers' Identities Catalog (& WikiData)

2015-12-15 Thread Gerard Meijssen 
Hoi,
Sorry, I understand sarcasm but I do not understand what it is based upon.
Thanks,
 GerardM

On 15 December 2015 at 20:10, John Erling Blad  wrote:

> There are some pretty good methods for optimizing the match process, but I
> have not seen any implementation for that against Wikidata items. Only
> things I've seen are some opportunistic methods. Duck tests gone wrong, or
> "Darn it was a platypus!"
>
> On Mon, Dec 14, 2015 at 11:19 PM, André Costa 
> wrote:
>
>> I'm planning to bring a few of the datasets into mix'n'match (@Magnus
>> this is the one I asked sbout on Twitter) in January but not all of them
>> are suitable and I believe separating KulturNav into multiple datasets on
>> mix'n'match maxes more sense and makes it more likely that they get matched.
>>
>> Some of the early adopters of KulturNav have been working with WMSE to
>> facilitate bi-directional matching. This is done on a dataset-by-dataset
>> level since different institutions are responsible for different datasets.
>> My hope is that mix'n'match will help in this area as well, even as a tool
>> for the institutions own staff who are often interested in matching entries
>> to Wikipedia (which most of the time means wikidata).
>>
>> @John: There are processes for matching kulturnav identifiers to wikidata
>> entities. Only afterwards are details imported. Mainly to source statements
>> [1] and [2]. There is some (not so user friendly) stats at [3].
>>
>> Cheers,
>> André
>>
>> [1]
>> https://www.wikidata.org/wiki/Wikidata:Requests_for_permissions/Bot/L_PBot_2
>> [2]
>> https://www.wikidata.org/wiki/Wikidata:Requests_for_permissions/Bot/L_PBot_3
>> [3] https://tools.wmflabs.org/lp-tools/misc/data/
>> --
>> André Costa
>> GLAM developer
>> Wikimedia Sverige
>>
>> Magnus Manske, 13/12/2015 11:24:
>>
>> >
>> > Since no one mentioned it, there is a tool to do the matching to WD much
>> > more efficiently:
>> > https://tools.wmflabs.org/mix-n-match/
>> 
>>
>> +1
>>
>> ___
>> 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
>
>
___
Wikidata mailing list
Wikidata@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata

Re: [Wikidata] Miga Classes and Properties Browser

2015-12-15 Thread Markus Krötzsch 

Hi,

Something to be noted here is that initial loading is quite a bit slower 
than it used to be, since there are a lot more classes now. We are 
looking into options of making this faster, but this might need a full 
rewrite to become really fast. The good thing is that loading only has 
to happen once per month (until the next data update).


There is a known issue with character encoding now, leading to "?" in 
some labels/descriptions. We are looking into it.


Another new feature is that we now also count properties used on 
property pages:


http://tools.wmflabs.org/wikidata-exports/miga/#_cat=Properties/Uses%20in%20properties=1%20-%202

Compared to the old data, we have a lot more objects in some classes now 
(it's amazing how many asteroids, first names, bands, and legal cases 
there are ...). It seems we have more than 10,000 galaxies on Wikidata 
already.


Here are our top-100 classes by number of instances:

http://tools.wmflabs.org/wikidata-exports/miga/#_cat=Classes/Number%20of%20direct%20instances=1%20-%201000

Other classes have a lot of subclasses rather than instances:

http://tools.wmflabs.org/wikidata-exports/miga/#_cat=Classes/Number%20of%20direct%20subclasses=1000%20-%2020

And of course, as usual, you can browse individual properties to see the 
classes of objects they are used on ("What kind of things have a 
diameter?") and browse classes to see which properties are typical for 
them ("What kind of statements do we have about poems?").


Cheers,

Markus


On 14.12.2015 22:00, Markus Damm wrote:

Hi all,

there  are some good news: I updated the Miga Classes and Properties
Browser which collects several statistics about classes and properties
used in Wikidata. In the future it will be updated monthly.

You can find it here: http://tools.wmflabs.org/wikidata-exports/miga/

Hint: Since Miga uses WebSQL, the browser does not run in Internet
Explorer or Mozilla Firefox.

Best regards,
Markus

___
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