1
00:00:06,589 --> 00:00:08,692
Alright. Let's do some presents.
2
00:00:09,659 --> 00:00:10,969
How this is going to work
3
00:00:10,969 --> 00:00:15,292
is that the presenters get
roughly two minutes
4
00:00:15,292 --> 00:00:19,431
to show their present.
5
00:00:21,107 --> 00:00:22,469
You get to applaud.
6
00:00:22,469 --> 00:00:25,224
Questions at the very, very end
if we have time.
7
00:00:27,488 --> 00:00:30,075
So, we're going to start
8
00:00:30,075 --> 00:00:31,620
with Joachim.
9
00:00:34,304 --> 00:00:35,563
Whos going to talk about
10
00:00:35,563 --> 00:00:37,843
the 20th Century Press Archives.
11
00:00:38,437 --> 00:00:39,519
Thanks.
12
00:00:42,491 --> 00:00:48,418
I'm presenting the first part
of a data donation by ZBW
13
00:00:48,418 --> 00:00:51,227
from the 20th Century Press Archives,
14
00:00:52,439 --> 00:00:54,517
which is to our best knowledge
15
00:00:54,517 --> 00:00:58,507
the largest public newspaper clippings
archived in the world.
16
00:01:00,011 --> 00:01:03,752
It has existed from 2008 to--
17
00:01:04,434 --> 00:01:07,066
from 1908 to 2005.
18
00:01:07,960 --> 00:01:12,158
And it evaluated more
than 1,500 periodicals
19
00:01:12,158 --> 00:01:15,565
from Germany and from all over the world.
20
00:01:17,325 --> 00:01:19,866
The material was organized
in folders
21
00:01:19,866 --> 00:01:21,107
as you see here.
22
00:01:22,337 --> 00:01:25,367
From a small corner
from a persons archive,
23
00:01:26,326 --> 00:01:29,231
and 25,000 folders
24
00:01:29,765 --> 00:01:31,896
with more than 2 million articles
25
00:01:31,896 --> 00:01:34,906
and digitalized pages
are online now.
26
00:01:37,459 --> 00:01:41,310
The integration of the persons archive
metadata to Wikidata
27
00:01:41,998 --> 00:01:44,336
has been completed most recently.
28
00:01:45,204 --> 00:01:47,983
And all of the more
than 5,000 person folders
29
00:01:47,983 --> 00:01:50,765
are accessible from Wikidata now.
30
00:01:52,140 --> 00:01:54,132
More than 6,000 facts
31
00:01:54,132 --> 00:01:57,188
sourced from the person's archive metadata
32
00:01:57,557 --> 00:01:59,259
has been edited Wikidata
33
00:01:59,259 --> 00:02:04,595
and this includes rather
complex relations
34
00:02:04,595 --> 00:02:08,269
like between persons and companies,
35
00:02:08,269 --> 00:02:10,702
and their role in the company.
36
00:02:13,560 --> 00:02:17,281
The next big challenge will be
37
00:02:17,281 --> 00:02:21,283
the countries and categories archive
38
00:02:21,705 --> 00:02:24,237
with more than 9,000 folders
39
00:02:24,851 --> 00:02:27,598
which is organized by--
40
00:02:28,286 --> 00:02:29,344
Yes?
41
00:02:29,715 --> 00:02:32,790
A hierarchy of countries
and hierarchy of categories.
42
00:02:32,790 --> 00:02:38,586
It's a whole system
of knowledge organization
43
00:02:39,445 --> 00:02:41,348
about the whole world.
44
00:02:41,536 --> 00:02:44,115
Materialized in newspaper clippings
45
00:02:44,115 --> 00:02:46,353
and to match this in data
is a challenge,
46
00:02:46,353 --> 00:02:49,352
so please consider to join
47
00:02:49,352 --> 00:02:52,569
the WikiProject
20th Century Press Archives.
48
00:02:52,744 --> 00:02:54,663
(applause)
49
00:02:58,516 --> 00:02:59,730
Thank you Joachim.
50
00:03:04,254 --> 00:03:05,417
Alright.
51
00:03:07,145 --> 00:03:08,373
Lucas.
52
00:03:08,568 --> 00:03:10,370
Stage is yours.
53
00:03:12,044 --> 00:03:13,091
Hello.
54
00:03:13,091 --> 00:03:14,661
I'm presenting two things,
55
00:03:14,661 --> 00:03:17,126
so I get four minutes--
I've been told I've hacked the system.
56
00:03:17,126 --> 00:03:18,615
(laughing)
So...
57
00:03:18,716 --> 00:03:19,784
The first thing is
58
00:03:19,784 --> 00:03:21,463
on behalf of Wikimedia Germany,
59
00:03:21,463 --> 00:03:25,958
which is first version
of Lua support for lexemes.
60
00:03:25,958 --> 00:03:27,142
(audience) Whooah!
61
00:03:27,592 --> 00:03:29,108
(loud applause)
62
00:03:31,242 --> 00:03:33,222
So you can see some Lua code here
63
00:03:33,222 --> 00:03:35,381
which is there's probably
not enough time to read that,
64
00:03:35,381 --> 00:03:37,101
and I'm not a great Lua programmer anyway,
65
00:03:37,101 --> 00:03:38,641
but the result is down there.
66
00:03:38,641 --> 00:03:41,438
We have access to the lexemes,
67
00:03:41,438 --> 00:03:42,460
forms, the census,
68
00:03:42,460 --> 00:03:45,023
also statements which are not
in the screenshot.
69
00:03:45,023 --> 00:03:47,436
And it's not deployed anywhere yet
70
00:03:47,436 --> 00:03:50,283
so that was just on my local Wiki.
(laughing)
71
00:03:50,660 --> 00:03:53,370
But we're hoping to get it
at least to beta soon.
72
00:03:53,370 --> 00:03:55,094
Probably to test WikiData
73
00:03:55,094 --> 00:03:56,690
pretty soon afterwards
74
00:03:56,690 --> 00:03:58,765
and then we'll see
where it goes,
75
00:03:58,765 --> 00:04:00,421
and it's a start at least.
76
00:04:00,825 --> 00:04:02,572
Thanks.
(applause)
77
00:04:05,326 --> 00:04:08,218
And the second thing
I'm doing as a volunteer
78
00:04:08,531 --> 00:04:09,623
so there's--
79
00:04:09,623 --> 00:04:12,403
I made this tool a while ago
called Wikidata Image Positions.
80
00:04:12,617 --> 00:04:16,642
So if you have a statement on item
that it depicts something,
81
00:04:16,642 --> 00:04:18,846
for example a paintinng
could depict a person,
82
00:04:18,846 --> 00:04:21,758
you can add a qualifier there
saying that this--
83
00:04:21,990 --> 00:04:24,921
where in the image this is so--
84
00:04:25,927 --> 00:04:27,784
like this person
is in the upper left corner
85
00:04:27,784 --> 00:04:28,895
of the image or something,
86
00:04:28,895 --> 00:04:32,398
and that is now supporting structured data
on Commons as well.
87
00:04:32,929 --> 00:04:36,358
And if the presents page is open
somewhere...
88
00:04:36,496 --> 00:04:38,707
No, not like that. I'm very sorry.
89
00:04:40,703 --> 00:04:43,677
We can change that.
90
00:04:43,937 --> 00:04:46,114
Yes, or read my emails.
(laughing)
91
00:04:47,699 --> 00:04:49,725
(audience 1) So much unread media.
92
00:04:49,725 --> 00:04:51,327
(laughing)
93
00:04:54,327 --> 00:04:56,142
- That one.
- There you go.
94
00:04:56,142 --> 00:04:58,240
So, there it is.
95
00:04:58,240 --> 00:05:01,251
This is going to be a picture I took
earlier this year,
96
00:05:01,251 --> 00:05:03,251
and there's already some
structured data here
97
00:05:03,251 --> 00:05:05,147
that says depicts
98
00:05:05,147 --> 00:05:06,536
certain pride flags,
99
00:05:06,536 --> 00:05:08,053
and once this loads,
100
00:05:08,053 --> 00:05:09,897
I can define the region.
101
00:05:10,325 --> 00:05:12,963
There we go,
and this also now used the same library
102
00:05:12,963 --> 00:05:17,236
as crop tool instead
of my home grown bad thing
103
00:05:17,236 --> 00:05:20,908
which I think Andy was
asking for years ago,
104
00:05:20,908 --> 00:05:22,332
and now it's finally done.
105
00:05:22,332 --> 00:05:23,697
And I say use this region.
106
00:05:23,697 --> 00:05:25,328
It's adding a qualifier.
107
00:05:25,328 --> 00:05:28,579
Let's do the same thing
over here.
108
00:05:28,702 --> 00:05:30,844
Just roughly drawn with a mouse.
109
00:05:30,844 --> 00:05:32,932
That should be good enough.
110
00:05:33,675 --> 00:05:35,395
And the third one.
111
00:05:37,648 --> 00:05:38,722
There we go.
112
00:05:38,722 --> 00:05:40,081
Use this region.
113
00:05:40,081 --> 00:05:42,897
And now if we check Lydia's contributions
114
00:05:42,897 --> 00:05:47,254
- except on Commons.
- (laughing)
115
00:05:48,329 --> 00:05:49,783
Media.org
116
00:05:49,783 --> 00:05:52,057
She gave me permission
to do this by the way.
117
00:05:52,057 --> 00:05:53,377
(laughing)
118
00:05:54,777 --> 00:05:56,387
User contirbutions.
119
00:05:56,777 --> 00:05:59,731
Where you can see
some new qualifiers here,
120
00:05:59,731 --> 00:06:00,924
and if we load this,
121
00:06:00,924 --> 00:06:03,238
there's also used script,
which is also hopefully working,
122
00:06:03,238 --> 00:06:06,036
- which shows you these regions correctly
- (crowd) Wooah! (applause)
123
00:06:06,307 --> 00:06:08,732
on Commons.
(applause)
124
00:06:12,192 --> 00:06:16,259
So basically the days of this
old annotation gadgets are numbered.
125
00:06:16,561 --> 00:06:17,678
(laughing)
126
00:06:18,698 --> 00:06:23,202
And that's it I think we can skip like
dozen back-up screenshots here now
127
00:06:23,358 --> 00:06:25,113
and go to the next person.
128
00:06:25,348 --> 00:06:28,176
- Thanks.
- (audience) Woohoo. (applause)
129
00:06:33,735 --> 00:06:35,003
Which one? This one.
130
00:06:35,217 --> 00:06:36,245
Okay,
131
00:06:36,245 --> 00:06:40,936
So we have a Lexeme
uploading bot.
132
00:06:41,962 --> 00:06:43,175
It's not developed by me.
133
00:06:43,175 --> 00:06:45,507
It's developed by Ehuyar Foundation.
134
00:06:46,071 --> 00:06:48,377
And it's developed to upload
135
00:06:48,377 --> 00:06:50,593
Basque language lexemes
with all its forms
136
00:06:50,593 --> 00:06:52,958
because its lexemes has 65 forms.
137
00:06:53,565 --> 00:06:56,266
So, it's not something
that we can do by hand easily.
138
00:06:56,852 --> 00:06:58,343
And also census.
139
00:07:01,285 --> 00:07:02,994
You can download there.
140
00:07:02,994 --> 00:07:04,577
I don't even know how it works.
141
00:07:04,577 --> 00:07:06,341
(laughing)
142
00:07:06,505 --> 00:07:09,783
And all I know is that
it's based in Wikidata Toolkit,
143
00:07:09,783 --> 00:07:13,430
so it's a subversion of that
and it's also in the [inaudible]
144
00:07:13,430 --> 00:07:15,198
of Wikidata, I think.
145
00:07:15,463 --> 00:07:16,698
That's it.
(applause)
146
00:07:16,698 --> 00:07:19,234
- (Lydia) Thank you.
- (audience) Woohoo (applause)
147
00:07:23,243 --> 00:07:24,494
Hello.
148
00:07:25,548 --> 00:07:27,821
I may have to reload this.
Let me just make sure.
149
00:07:32,750 --> 00:07:35,318
Does it work? Yes. Present.
150
00:07:36,164 --> 00:07:37,212
Great.
151
00:07:37,212 --> 00:07:39,655
Now, this is a project
we've been working on for a while,
152
00:07:39,655 --> 00:07:41,480
but we are rolling it out
for the first time
153
00:07:41,480 --> 00:07:43,935
to a big crowd here
at WIkidataCon
154
00:07:43,935 --> 00:07:46,346
and we'll show you some--
a real cool feature
155
00:07:46,346 --> 00:07:48,149
that we didn't tell you about
earlier today.
156
00:07:48,149 --> 00:07:51,398
So with this is a project called
The Wiki Art Depiction Explorer,
157
00:07:51,398 --> 00:07:54,137
and this is an attempt
to try give an interface
158
00:07:54,137 --> 00:07:56,711
better than
just editing a raw Wikidata item
159
00:07:56,711 --> 00:07:58,970
when it comes
to adding depiction information
160
00:07:58,970 --> 00:08:00,930
for artworks,
so this is a project that was funded
161
00:08:00,930 --> 00:08:04,050
by the Night Foundation,
and Wikimedia D.C.
162
00:08:04,050 --> 00:08:06,431
and the Smithsonian,
we worked together on this project
163
00:08:06,431 --> 00:08:08,275
and with the amazing development skills
164
00:08:08,275 --> 00:08:10,424
of Edward Betts, right here.
165
00:08:11,066 --> 00:08:12,970
So this is an example
of what you'll see,
166
00:08:12,970 --> 00:08:14,541
and we invite you all to try it out.
167
00:08:14,541 --> 00:08:16,998
art.wikidata.link
is the URL,
168
00:08:16,998 --> 00:08:21,017
and the idea here is that
you can see a large version of the picture
169
00:08:21,017 --> 00:08:23,314
and we will try to bring in
whatever we can
170
00:08:23,314 --> 00:08:26,169
from the object page
of the institution
171
00:08:26,169 --> 00:08:27,798
that holds this image.
172
00:08:27,798 --> 00:08:30,573
So, we're bringing
in some description information,
173
00:08:30,573 --> 00:08:32,926
so that the person trying
to add depict information,
174
00:08:32,926 --> 00:08:36,297
has some additional readings
that they can have here,
175
00:08:36,297 --> 00:08:39,500
We also bring in some key words,
and the great thing about this
176
00:08:39,500 --> 00:08:41,695
when you start typing in the box,
177
00:08:41,695 --> 00:08:45,208
you are actually given matches
against whatever's been previously
178
00:08:45,208 --> 00:08:48,418
matched in depictions statements.
179
00:08:48,797 --> 00:08:51,089
So, this is a much
tighter controlled vocabulary.
180
00:08:51,455 --> 00:08:54,095
It gives you a much better direction
of what to do.
181
00:08:54,095 --> 00:08:56,999
So here is an example here,
if this works correctly,
182
00:08:56,999 --> 00:08:58,448
we should be able to click that.
183
00:08:58,448 --> 00:09:01,600
And also make more edits
on Lydia's behalf. (laughing)
184
00:09:01,600 --> 00:09:03,380
And we can go in here
and type in ballet,
185
00:09:03,380 --> 00:09:05,780
and you'll see that it doesn't match
everything on Wikidata
186
00:09:05,780 --> 00:09:07,551
but only things that are relevant
187
00:09:07,551 --> 00:09:09,811
based on previous depiction,
statements,
188
00:09:09,811 --> 00:09:11,282
so I can say ballet dancer.
189
00:09:11,282 --> 00:09:15,700
I can go back in there
and add these different--
190
00:09:15,912 --> 00:09:17,151
Oops.
191
00:09:19,931 --> 00:09:22,141
Ooh. Not sure
why it's not working.
192
00:09:22,141 --> 00:09:24,135
Anyway, let me go ahead
and make those edits
193
00:09:24,135 --> 00:09:26,560
and then now that has been committed
194
00:09:26,560 --> 00:09:28,559
and you can actually start
browsing other things.
195
00:09:28,559 --> 00:09:31,245
So the idea is to keep you
in this universe of paintings
196
00:09:31,491 --> 00:09:34,502
and artworks, and not just
punch you back out to Wikidata.
197
00:09:34,817 --> 00:09:38,172
But we also have another
bonus function
198
00:09:38,172 --> 00:09:40,640
that Edward is going to show you.
199
00:09:46,270 --> 00:09:52,214
So, this painting we've got date of birth
and death for this person.
200
00:09:52,415 --> 00:09:56,442
So, we do a search
and these are all the people
201
00:09:56,442 --> 00:09:59,411
that were born and died in those years
202
00:09:59,786 --> 00:10:00,830
and so...
203
00:10:02,372 --> 00:10:03,503
are we at Aurora?
204
00:10:03,503 --> 00:10:05,675
- Yep. Can you see the match?
- Which one is it?
205
00:10:05,675 --> 00:10:07,516
- I didn't see the match.
- Yep, further down.
206
00:10:07,516 --> 00:10:10,011
Maria Aurora. Right there.
207
00:10:10,245 --> 00:10:12,247
- So, if I click on that.
- So, you click on that.
208
00:10:12,247 --> 00:10:13,502
and then scroll down.
209
00:10:13,502 --> 00:10:15,571
And then we go-- Oh it's...
(laughing)
210
00:10:16,126 --> 00:10:17,487
There you go.
211
00:10:17,647 --> 00:10:18,965
Add these to the painting.
212
00:10:18,965 --> 00:10:22,269
(audience 2) You need to highlight
word metrics from the title.
213
00:10:22,282 --> 00:10:24,230
That... it'll come.
(laughing)
214
00:10:24,480 --> 00:10:25,811
So, it worked.
215
00:10:25,811 --> 00:10:27,217
It saved it to the painting.
216
00:10:27,217 --> 00:10:30,216
So it'll match all humans
with those birth date
217
00:10:30,216 --> 00:10:31,352
and the death date,
218
00:10:31,352 --> 00:10:33,393
and you can click that automatically.
219
00:10:33,393 --> 00:10:37,405
- And that's it. So go ahead and try it.
- (audience) Woohoo! (applause)
220
00:10:39,069 --> 00:10:41,170
Oh, those are just some stats.
221
00:10:41,170 --> 00:10:43,438
We had a whole bunch
of people try it already today,
222
00:10:43,438 --> 00:10:45,228
and we upped this number today.
223
00:10:45,228 --> 00:10:47,373
So keep working at it. Thanks.
224
00:10:57,548 --> 00:10:58,907
Next one is Bruno.
225
00:11:04,825 --> 00:11:05,911
Hello.
226
00:11:05,911 --> 00:11:07,407
I'm Bruno from Google
227
00:11:07,407 --> 00:11:09,210
and we are open sourcing
228
00:11:09,210 --> 00:11:11,137
what we call lexical masks.
229
00:11:11,137 --> 00:11:13,797
A Little bit less sexier than the picture
we just saw.
230
00:11:14,724 --> 00:11:19,393
It's a config file that specifies
what a lexeme needs to look like,
231
00:11:19,393 --> 00:11:21,973
what kind of form,
you expect in a lexeme
232
00:11:21,973 --> 00:11:24,557
and what kind of feature you want
on those forms.
233
00:11:25,029 --> 00:11:27,631
Example here, it's German nouns
234
00:11:27,631 --> 00:11:30,182
that will have a gender inherent,
235
00:11:30,182 --> 00:11:33,452
and we'll have a couple of forms
specified here
236
00:11:33,452 --> 00:11:35,825
with a couple of features you expect.
237
00:11:36,396 --> 00:11:40,171
The mask or the text files
that you see here
238
00:11:40,171 --> 00:11:44,313
will be uploaded to Wikidata
so that I can help the Lexeme community
239
00:11:44,313 --> 00:11:48,315
to check consistency
and increase the coverage.
240
00:11:48,315 --> 00:11:51,010
More details on the talk I gave
earlier today
241
00:11:51,010 --> 00:11:52,347
and Lydia's.
242
00:11:52,347 --> 00:11:54,837
Thank you.
(applause)
243
00:12:03,024 --> 00:12:05,940
Yes, so in the past two years,
244
00:12:05,940 --> 00:12:07,705
I have had an hobby
245
00:12:07,705 --> 00:12:09,593
because I was not, let's say,
very happy
246
00:12:09,593 --> 00:12:12,727
with the current SPARQL implementation
especially Blazegraph.
247
00:12:13,016 --> 00:12:15,611
So, during my free time,
I started a project
248
00:12:15,611 --> 00:12:16,868
I called Oxigraph
249
00:12:16,868 --> 00:12:20,233
so, it's basically like Blazegraph
but different.
250
00:12:20,233 --> 00:12:21,698
(laughing)
251
00:12:21,805 --> 00:12:24,363
So, it starts getting in those
states that work
252
00:12:24,363 --> 00:12:26,665
so SPARQL queries were implemented.
253
00:12:26,774 --> 00:12:29,454
But it's not been optimized yet,
currently there is no optimization
254
00:12:29,454 --> 00:12:31,207
of how queries are executed.
255
00:12:31,207 --> 00:12:33,831
So as you're seeing
this small experiment
256
00:12:33,831 --> 00:12:37,478
with some SPARQL queries
the results seem fairly promising.
257
00:12:38,092 --> 00:12:40,450
What is nice is I used the rest
to implement it
258
00:12:40,450 --> 00:12:44,526
and I managed to get the memory footprint
fairly reasonable
259
00:12:44,526 --> 00:12:47,542
as well as some--
Blazegraph origin
260
00:12:47,542 --> 00:12:50,926
or [inaudible] so,
I hope that maybe in the future
261
00:12:50,926 --> 00:12:54,217
I'm going to get
maybe all the people who can
262
00:12:54,217 --> 00:12:56,913
then spend more time to make it
ready working well,
263
00:12:56,913 --> 00:13:01,157
we could have something very good
for at least smaller Wikibase [inaudible]
264
00:13:01,157 --> 00:13:04,313
with a few million
or ten of million [inaudible].
265
00:13:04,758 --> 00:13:09,254
So the repository is here,
so it's not working fairly well.
266
00:13:09,254 --> 00:13:12,823
It's a work in progress
and if you want to test it,
267
00:13:12,823 --> 00:13:14,524
or contribute,
you are much welcome
268
00:13:14,524 --> 00:13:16,224
because it's a big task.
269
00:13:16,224 --> 00:13:17,350
Thank you.
270
00:13:17,350 --> 00:13:19,630
(audience cheers and applause)
271
00:13:29,422 --> 00:13:30,769
I was going to do a live demo
272
00:13:30,769 --> 00:13:32,005
but it didn't go well earlier,
273
00:13:32,005 --> 00:13:33,340
so this is a video,
(laughing)
274
00:13:33,340 --> 00:13:34,791
which...
275
00:13:35,544 --> 00:13:38,179
Wait, I can do this on my phone right?
276
00:13:39,839 --> 00:13:41,157
Access denied.
277
00:13:42,231 --> 00:13:44,298
(singing) Ta-ta-ta-ta-ta
278
00:13:45,971 --> 00:13:46,988
Two minutes.
279
00:13:46,988 --> 00:13:49,534
Well the video is two minutes long.
(laughing)
280
00:13:51,167 --> 00:13:52,458
Ta-ta-ta-ta-ta
281
00:13:53,865 --> 00:13:55,934
(audience 3) We have two people here
from Google.
282
00:13:55,934 --> 00:13:58,821
I'm sure they can come up
with something. (laughing)
283
00:14:00,311 --> 00:14:02,802
Okay, we might just watch
it in this tab.
284
00:14:03,166 --> 00:14:04,226
Come on.
285
00:14:04,226 --> 00:14:05,433
Da-da-da-da-da
286
00:14:05,433 --> 00:14:07,926
Yeah, so I did a live demo
and I thought that would go badly.
287
00:14:07,926 --> 00:14:09,235
But--
(laughing)
288
00:14:09,554 --> 00:14:12,491
So, this is something I've been working on
289
00:14:12,491 --> 00:14:14,885
ever since creating
the doc images two years ago
290
00:14:14,885 --> 00:14:18,826
and this is a sort of
shared platform website
291
00:14:18,826 --> 00:14:20,729
where you can go ahead
and make an account.
292
00:14:21,315 --> 00:14:22,763
Ta-ta-ta-ta-ta
293
00:14:23,151 --> 00:14:25,354
And then you've got a lovely button
on the next page
294
00:14:25,354 --> 00:14:27,414
which allows you to create a Wiki.
295
00:14:27,461 --> 00:14:29,429
There's lot of features
missing at the moment,
296
00:14:29,429 --> 00:14:30,842
so you get to choose a Wiki name,
297
00:14:30,842 --> 00:14:32,897
where it is and a user name at the moment.
298
00:14:33,520 --> 00:14:35,706
But the possibilities are endless,
299
00:14:35,706 --> 00:14:38,777
and it goes and creates a Wiki
in a shared enviroment,
300
00:14:39,321 --> 00:14:41,726
saving on all of us
expensive resources
301
00:14:41,726 --> 00:14:44,240
that we're all spending
running Wikibases.
302
00:14:46,074 --> 00:14:48,244
I recorded this just downstairs earlier
303
00:14:48,244 --> 00:14:50,746
so, this is like, kind of real.
304
00:14:50,746 --> 00:14:53,547
I sped it up slightly but--
305
00:14:53,864 --> 00:14:57,409
You get emailed your media Wiki
temporary password.
306
00:14:57,409 --> 00:14:59,772
You can log in with the user account
that you made.
307
00:15:00,084 --> 00:15:01,745
Da-da-da-da.
308
00:15:02,595 --> 00:15:04,190
Then you have to change your password.
309
00:15:04,190 --> 00:15:06,554
This is where I just copy
part of the URL in.
310
00:15:06,554 --> 00:15:08,288
(laughing)
311
00:15:10,435 --> 00:15:13,029
And then you're logged into your
very own Wikibase
312
00:15:13,029 --> 00:15:15,745
that has quick statements,
a career service,
313
00:15:15,745 --> 00:15:18,839
everything managed for you
that you don't have to worry about.
314
00:15:18,947 --> 00:15:22,004
(audience) Woo! Woo woo woo.
(applauding)
315
00:15:26,339 --> 00:15:27,553
So you can go and create
316
00:15:27,553 --> 00:15:30,109
all of your items, use tools,
317
00:15:30,109 --> 00:15:32,939
and I plan on adding
more tools in the future.
318
00:15:33,518 --> 00:15:35,500
All of the complexities are hidden.
319
00:15:35,500 --> 00:15:37,324
At the moment this is live
320
00:15:37,324 --> 00:15:40,048
on this wbstat.com domain name.
321
00:15:40,879 --> 00:15:42,756
But you need an invitation code from me.
322
00:15:42,756 --> 00:15:45,074
If you want one to try one out
during WikidataCon,
323
00:15:45,074 --> 00:15:47,791
come and talk to me,
and I will give you one.
324
00:15:47,791 --> 00:15:51,165
It's full of bugs at the moment,
and stuff so don't rely on it.
325
00:15:52,854 --> 00:15:54,398
This is quick statements working,
326
00:15:54,398 --> 00:15:56,348
and then on the Saturday
of WikidataCon
327
00:15:56,348 --> 00:15:57,829
I'll delete all of the data,
328
00:15:57,829 --> 00:16:00,567
and then we'll be using this
for the Wikibase workshops
329
00:16:00,567 --> 00:16:02,951
on Sunday if any of you are attending.
330
00:16:03,682 --> 00:16:05,638
And then it will get a real test,
331
00:16:05,638 --> 00:16:07,943
and so then you can see
the two edits have happened.
332
00:16:07,943 --> 00:16:10,462
I'm so glad I didn't do this
as live demo,
333
00:16:10,462 --> 00:16:13,091
and then you go to the query service.
334
00:16:13,091 --> 00:16:16,210
You type in the query
that shows you all of the triples.
335
00:16:17,713 --> 00:16:20,351
Even in the recording, you do it wrong.
(laughing)
336
00:16:20,351 --> 00:16:21,779
And then you have all your triples.
337
00:16:21,779 --> 00:16:24,039
(audiences cheers and applause)
338
00:16:36,571 --> 00:16:37,841
Alright.
339
00:16:41,957 --> 00:16:44,218
So, Happy Birthday Wikidata!
340
00:16:45,238 --> 00:16:47,069
I started on this last year,
341
00:16:47,069 --> 00:16:48,772
actually for the Wiki site meeting
342
00:16:48,772 --> 00:16:51,420
which was about a year ago, and--
343
00:16:51,937 --> 00:16:54,639
got something running
and got a lot encouragement
344
00:16:54,639 --> 00:16:57,157
from Daniel Mietchen
who's probably watching online.
345
00:16:57,157 --> 00:16:58,959
Hi Daniel.
(laughing)
346
00:17:00,574 --> 00:17:02,846
And he's given me all sorts of ideas
347
00:17:02,846 --> 00:17:06,281
for improving it,
and so just in time for this meeting,
348
00:17:06,622 --> 00:17:09,539
I've got a new version out
that does more.
349
00:17:09,857 --> 00:17:12,893
So basically what this is
is a tool to replace
350
00:17:12,893 --> 00:17:16,683
in mostly scientific articles
but any work really
351
00:17:16,683 --> 00:17:19,959
where there's an author--
author name string,
352
00:17:19,959 --> 00:17:22,974
replace that with an actual
author item which is--
353
00:17:22,974 --> 00:17:26,404
pulled from Wikidata using
some various things to match
354
00:17:26,404 --> 00:17:28,085
to that properly.
355
00:17:28,085 --> 00:17:31,827
So changes are most recently
356
00:17:31,827 --> 00:17:34,040
you can log in with your Wikimedia
user account
357
00:17:34,040 --> 00:17:37,055
and do the edits directly
rather than previously
358
00:17:37,055 --> 00:17:38,727
it all just went through quick statements
359
00:17:38,727 --> 00:17:41,996
which I've got about a million
quick statements edits now.
360
00:17:41,996 --> 00:17:43,153
laughing)
361
00:17:43,153 --> 00:17:45,022
This is a llittle bit past here
362
00:17:45,022 --> 00:17:47,187
and bypasses that.
363
00:17:48,238 --> 00:17:49,988
Another new thing
is you can actually go in
364
00:17:49,988 --> 00:17:51,373
and look at a work,
365
00:17:51,373 --> 00:17:54,510
and update all of the authors
on that work at once.
366
00:17:54,510 --> 00:17:56,043
So there's a match button.
367
00:17:56,043 --> 00:18:00,031
You can also rearrange the author list
if they're out of order or something.
368
00:18:01,882 --> 00:18:03,271
There's also---
369
00:18:04,497 --> 00:18:06,583
Oh yeah, this is an example
of what that looks like
370
00:18:06,583 --> 00:18:09,302
so when you're matching
it up with authors it lists...
371
00:18:10,297 --> 00:18:13,598
some information about their affliation
as it is in Wikidata.
372
00:18:15,025 --> 00:18:18,653
The other thing that's new
is some automatic filtering.
373
00:18:18,653 --> 00:18:21,135
If you go to the bottom of a page
374
00:18:21,135 --> 00:18:23,768
that is a search for an author name,
375
00:18:24,052 --> 00:18:25,714
you'll see links to coauthors,
376
00:18:25,714 --> 00:18:28,111
links to other--
to the topics they've written on,
377
00:18:28,111 --> 00:18:29,456
links to their journals
378
00:18:29,456 --> 00:18:32,038
and so you can filter,
and narrow down the list
379
00:18:32,038 --> 00:18:33,877
of your listed works
that you're looking at
380
00:18:33,877 --> 00:18:35,128
to just those things
381
00:18:35,128 --> 00:18:37,184
that have those particular
features in them.
382
00:18:37,184 --> 00:18:38,570
Anyway, that's what's new there.
383
00:18:38,570 --> 00:18:40,758
And that's it. Thank you.
384
00:18:40,758 --> 00:18:43,279
(cheers and applause)
385
00:18:55,608 --> 00:18:56,673
Hi all.
386
00:18:57,454 --> 00:18:59,794
I'm also from WMDE
387
00:18:59,794 --> 00:19:02,385
and a volunteer,
but this is volunteer work.
388
00:19:04,114 --> 00:19:05,826
It's a tool MachtSinn.
389
00:19:05,826 --> 00:19:09,690
A few of them might...
a few might already seen it,
390
00:19:10,319 --> 00:19:14,494
but I improved it greatly
in the last week.
391
00:19:15,354 --> 00:19:18,284
So we have these lexemes nowadays
392
00:19:18,284 --> 00:19:22,611
and on these lexemes
you should add the sense
393
00:19:22,611 --> 00:19:26,328
what the word means,
and all the different meanings
394
00:19:26,328 --> 00:19:27,916
a word can have,
395
00:19:27,916 --> 00:19:33,581
and we have a lot of lexemes now
that still are missing senses,
396
00:19:34,028 --> 00:19:38,079
that don't have any sense
and... (laughing)
397
00:19:38,688 --> 00:19:43,420
in a lot of cases,
we also have items
398
00:19:43,820 --> 00:19:48,326
about the concept that
this sense is about,
399
00:19:48,585 --> 00:19:52,701
and so I thought we could mind
the senses,
400
00:19:52,701 --> 00:19:56,048
and this is what MatchtSinn does.
401
00:19:56,833 --> 00:19:59,834
It shows you for a lexeme
402
00:19:59,834 --> 00:20:03,437
and this case the English
word tune,
403
00:20:03,437 --> 00:20:07,482
which is a verb
and possible meaning.
404
00:20:07,868 --> 00:20:10,744
In this case: short instrumental piece,
405
00:20:11,187 --> 00:20:13,668
a melody and you're asked
406
00:20:13,668 --> 00:20:17,243
is this a meaning for this word,
407
00:20:17,243 --> 00:20:19,820
and if you click the blue button
408
00:20:19,820 --> 00:20:22,063
it will save it to Wikidata.
409
00:20:22,816 --> 00:20:24,767
And if you click the right button,
410
00:20:24,767 --> 00:20:27,206
it will throw it away.
411
00:20:28,062 --> 00:20:31,933
And you login to--
412
00:20:31,933 --> 00:20:34,379
the tool with your Wikidata account.
413
00:20:34,379 --> 00:20:35,814
We are OAuth.
414
00:20:35,814 --> 00:20:38,248
And a few people have already using it
415
00:20:38,248 --> 00:20:41,882
and added 6,000 senses,
416
00:20:42,308 --> 00:20:45,929
and there are currently
about 40,000 senses
417
00:20:45,929 --> 00:20:49,979
waiting to be considered,
418
00:20:49,979 --> 00:20:54,554
and tested,
and also for writing this
419
00:20:54,554 --> 00:21:00,282
I had to first write some Python tool
420
00:21:00,282 --> 00:21:05,073
to modify Lexemes because
Pi Wiki bot
421
00:21:05,073 --> 00:21:10,289
and the other common tools
don't support that.
422
00:21:10,716 --> 00:21:11,912
Yes, thanks.
423
00:21:12,537 --> 00:21:14,305
(applause)
424
00:21:27,494 --> 00:21:29,750
Hi, Happy Birthday.
(laughing)
425
00:21:30,334 --> 00:21:31,469
Happy Birthday!
426
00:21:31,952 --> 00:21:33,897
(audience) Woo!
(laughing)
427
00:21:34,640 --> 00:21:36,806
Are you eating cake
during my presentation?
428
00:21:36,806 --> 00:21:41,553
Okay, what to expect from a Data Scientist
for [inaudible] dashboard, of course.
429
00:21:42,098 --> 00:21:43,531
Okay, so this time--
430
00:21:44,565 --> 00:21:46,335
well, there's the end product. (laughing)
431
00:21:46,901 --> 00:21:49,534
This time something called
Wikidata Languages Landscape.
432
00:21:50,244 --> 00:21:52,515
So--it's is a dashboard
as I said
433
00:21:52,515 --> 00:21:55,199
so some of the empirical findings
that you present
434
00:21:55,199 --> 00:21:58,881
through the Wikidata statistics
were already serviced today
435
00:21:58,881 --> 00:22:01,883
in Lydia's talks,
and basically focuses
436
00:22:01,883 --> 00:22:04,687
on the structural organization
of languages in Wikidata,
437
00:22:04,687 --> 00:22:06,838
on the similarity of Wikidata languages
438
00:22:06,838 --> 00:22:11,768
in respect how they're reused across
the Wikimedia Foundation projects, right?
439
00:22:12,373 --> 00:22:15,729
And it also combines
some of the external resources
440
00:22:15,729 --> 00:22:19,557
with those statistics in order
to provide for a comprehensive view
441
00:22:19,557 --> 00:22:24,221
of how different languages cope
in this Wikimedia universe.
442
00:22:24,221 --> 00:22:26,535
So, there's a link to the dashboard
443
00:22:26,535 --> 00:22:30,483
so I was warned not to do this,
but I will try. (laughing)
444
00:22:30,483 --> 00:22:31,843
Sorry.
(laughing)
445
00:22:33,061 --> 00:22:38,608
So depending on the--
Yes. Yes! It can be done. (laughing)
446
00:22:38,836 --> 00:22:43,497
Okay, so I will be even able
to do a live demo maybe...
447
00:22:43,497 --> 00:22:45,376
Okay, come on, come on, come on.
448
00:22:45,376 --> 00:22:47,925
It's still computing.
It's a very complicated service.
449
00:22:48,400 --> 00:22:49,536
Yep, here we go.
450
00:22:49,536 --> 00:22:56,024
Okay, the first thing
that you will be able to see.
451
00:22:56,513 --> 00:22:57,819
Okay. Yes alright.
452
00:22:57,819 --> 00:22:58,920
(audience) Wow...
453
00:22:58,920 --> 00:23:02,300
Wow. As I said a Data Scientist,
so this is not really informative, right?
454
00:23:02,300 --> 00:23:03,318
Okay.
455
00:23:03,318 --> 00:23:07,652
So here you have all the languages,
or most of languages in WikiData
456
00:23:07,652 --> 00:23:09,539
and we're focusing on
those languages
457
00:23:09,539 --> 00:23:12,829
that were ever used anywhere
in Wikimedia, okay?
458
00:23:12,829 --> 00:23:16,253
So, we're talking about languages
that actually have labels for things
459
00:23:16,253 --> 00:23:19,109
that are mentioned in--
460
00:23:19,109 --> 00:23:21,617
across Wikipedia, Wikivoyage
and other projects, right?
461
00:23:21,617 --> 00:23:26,318
So, and this is only
a subontology of languages like so.
462
00:23:26,318 --> 00:23:31,677
This depicts only the instance
of [inaudible] in [inaudible]
463
00:23:31,677 --> 00:23:34,278
So that you can actually use
probably these tools here
464
00:23:34,278 --> 00:23:35,661
to browse this thing.
465
00:23:35,661 --> 00:23:38,302
It's not aesthetically pleasing
but at least it's complete.
466
00:23:39,007 --> 00:23:42,197
One of the byproducts
of this work is--
467
00:23:42,370 --> 00:23:45,571
sorry, not this thing,
but this thing here.
468
00:23:45,863 --> 00:23:48,169
So, this is a small visual browser
469
00:23:48,169 --> 00:23:51,764
that can help you
figure out what is wrong
470
00:23:51,764 --> 00:23:53,506
with the ontology of languages
in Wikidata,
471
00:23:53,506 --> 00:23:56,706
and if you want to fix something
it makes it easier for you to find.
472
00:23:57,110 --> 00:23:58,538
So, while working on this thing.
473
00:23:58,538 --> 00:24:01,729
I figured out that the languages
ontology's particularly complex,
474
00:24:01,729 --> 00:24:03,052
really complicated, okay.
475
00:24:03,052 --> 00:24:06,245
And then there's some inconsistency
there for example.
476
00:24:06,741 --> 00:24:09,106
Well at least in my intuitive
understanding of semantics,
477
00:24:09,106 --> 00:24:11,575
you can't be at the same time
a part of something
478
00:24:11,575 --> 00:24:12,717
in a subclass of something.
479
00:24:12,717 --> 00:24:15,252
I mean you can,
and Wikidata is really flexible enough
480
00:24:15,252 --> 00:24:16,558
to allow you to do that,
481
00:24:16,558 --> 00:24:19,339
but probably some things
need fixing in that respect,
482
00:24:19,339 --> 00:24:20,372
and here for example,
483
00:24:20,372 --> 00:24:21,518
you can find the language,
484
00:24:21,518 --> 00:24:23,026
say, for example, Serbo-Croation.
485
00:24:23,026 --> 00:24:24,695
It used to be my native language
486
00:24:24,695 --> 00:24:27,797
before it fell apart
into Serbian, Croatian, Bosnian, etc.
487
00:24:27,797 --> 00:24:30,337
Okay, and here you have
all the relations
488
00:24:30,337 --> 00:24:33,165
like P31 part of subclass [inaudible]
different marks.
489
00:24:33,165 --> 00:24:37,788
So, if anything needs to be fixed
instead of browsing the whole--
490
00:24:38,349 --> 00:24:40,908
the whole structure
of the whole ontology
491
00:24:40,908 --> 00:24:43,606
you can go here
and just make it shorter, right?
492
00:24:43,986 --> 00:24:46,390
And then things like plastering
the language
493
00:24:46,390 --> 00:24:48,035
many people like this on Twitter.
494
00:24:48,035 --> 00:24:50,895
It actually cost me
half of my life
495
00:24:50,895 --> 00:24:52,144
to produce this thing.
496
00:24:52,144 --> 00:24:53,547
Okay, it's huge.
497
00:24:53,547 --> 00:24:55,481
So yeah, this is the dashboard
498
00:24:55,481 --> 00:24:57,894
to go play
many interesting things.
499
00:24:57,894 --> 00:25:00,110
Thank you very much.
500
00:25:00,110 --> 00:25:01,655
(applause)
501
00:25:15,670 --> 00:25:17,350
So, this is--
502
00:25:17,674 --> 00:25:20,567
This presenation is called
WikiShape
503
00:25:20,567 --> 00:25:22,585
and this is something
that we have been working
504
00:25:22,585 --> 00:25:25,102
in the Shape Expressions community group,
505
00:25:25,552 --> 00:25:29,580
and the idea of WikiShape is that
506
00:25:29,580 --> 00:25:33,400
we want to have like
the Wikidata query service
507
00:25:33,400 --> 00:25:36,186
which is I think is something
508
00:25:36,186 --> 00:25:38,663
that most of you are using
to do SPARQL queries,
509
00:25:38,663 --> 00:25:41,107
but now we want to do
the same thing
510
00:25:41,107 --> 00:25:42,846
but for Shape Expressions.
511
00:25:42,846 --> 00:25:45,525
So we wanted an editor
512
00:25:45,525 --> 00:25:49,675
which is as easy to do,
and to work with it
513
00:25:49,675 --> 00:25:53,297
as it is with Wikidata query service.
514
00:25:53,297 --> 00:25:55,542
That's what we are going to--
515
00:25:55,802 --> 00:25:57,577
that's the world of WikiShape
516
00:25:57,577 --> 00:26:00,962
so you have Shape Expressions
editor and validator.
517
00:26:00,962 --> 00:26:03,027
You also have syntax highlighting,
518
00:26:03,027 --> 00:26:05,797
auto-completion,
schema visualiztion,
519
00:26:05,797 --> 00:26:07,136
and search.
520
00:26:07,136 --> 00:26:08,741
So this is just a screen.
521
00:26:09,759 --> 00:26:11,557
Well, this is--
I could click on that
522
00:26:11,557 --> 00:26:13,567
but I prefer not to do that.
523
00:26:14,046 --> 00:26:16,177
You can have info about schema.
524
00:26:16,177 --> 00:26:18,929
You can have sign information
about the schema.
525
00:26:18,929 --> 00:26:20,623
You can visualize the schema.
526
00:26:21,045 --> 00:26:23,234
This is--you can autocomplete.
527
00:26:23,729 --> 00:26:26,079
For example just start writing work
528
00:26:26,079 --> 00:26:28,753
and it finds the schemas for that.
529
00:26:28,900 --> 00:26:32,631
Then, you also have the editor
530
00:26:32,631 --> 00:26:35,250
and as you can see,
this is for written work.
531
00:26:35,250 --> 00:26:39,539
You can have this editor
of the schema of the Shape Expression.
532
00:26:40,305 --> 00:26:42,585
If you hover with a mouse,
533
00:26:43,285 --> 00:26:49,193
it highlights the name of the label
of the property
534
00:26:49,481 --> 00:26:51,770
which is the same
as the Wikidata Query service
535
00:26:51,770 --> 00:26:54,587
so the goal is that
536
00:26:54,587 --> 00:26:56,493
now that you have Shape Expressions,
537
00:26:56,493 --> 00:26:59,754
the goal is that you are
using Shape Expressions
538
00:26:59,754 --> 00:27:01,714
to validate your data
to increase the quality
539
00:27:01,714 --> 00:27:05,666
of the Wikidata--data...
540
00:27:06,187 --> 00:27:08,067
using Shape expressions.
541
00:27:08,148 --> 00:27:12,469
And also, you also can visualize
the schemas.
542
00:27:12,612 --> 00:27:15,389
Once you have a schema
for example for written work,
543
00:27:15,389 --> 00:27:17,238
the author or whatever
you can visualize,
544
00:27:17,238 --> 00:27:18,453
and all that.
545
00:27:18,453 --> 00:27:20,718
So, that's the goal of WikiShape.
546
00:27:20,718 --> 00:27:22,565
(applause)
547
00:27:37,428 --> 00:27:39,000
I love Wikidata.
548
00:27:39,057 --> 00:27:42,426
I'm very proud of the work I do
and my friends do on Wikidata,
549
00:27:42,760 --> 00:27:46,255
and I know most of you are pleased
to work on Wikidata as well.
550
00:27:46,601 --> 00:27:48,763
It's come to my attention
over the last couple of days
551
00:27:48,763 --> 00:27:50,165
that a couple of you
552
00:27:50,165 --> 00:27:52,072
are working on a rival product
(laughing)
553
00:27:52,072 --> 00:27:54,567
and undermining what happens
on Wikidata.
554
00:27:54,567 --> 00:27:57,432
This product is apparently called,
"Wiki-dah-ta" (enunciates the 'a')
555
00:27:57,432 --> 00:27:58,725
(laughing)
556
00:27:58,725 --> 00:28:01,229
I've never heard
of this "Wikidata" before.
557
00:28:02,124 --> 00:28:03,789
So, in order to get things correct
558
00:28:03,789 --> 00:28:06,401
because some of you are doing it wrong.
(laughing)
559
00:28:06,401 --> 00:28:08,919
If I can find the mouse pointer,
how do I open this?
560
00:28:09,527 --> 00:28:10,925
Here we go.
561
00:28:19,160 --> 00:28:21,029
We have here.
(laughing)
562
00:28:21,221 --> 00:28:22,441
(computer) Wikidata.
563
00:28:22,614 --> 00:28:23,652
You probably can't hear.
564
00:28:23,652 --> 00:28:26,118
That is very, very quiet.
Let me do that again.
565
00:28:26,305 --> 00:28:28,403
- (audience) No the other one.
- The one next to it.
566
00:28:28,403 --> 00:28:29,894
(Andy) Oh yep, I got you.
567
00:28:31,079 --> 00:28:33,034
You're not going to hear this anyway.
568
00:28:37,176 --> 00:28:39,270
(computer voice) Wiki-day-ta.
(laughing)
569
00:28:39,270 --> 00:28:41,735
So, there it is for the record.
(applause)
570
00:28:41,735 --> 00:28:44,666
But, okay.
571
00:28:48,009 --> 00:28:50,900
Joking aside if those of you
who do have speech impediments
572
00:28:50,900 --> 00:28:53,091
would like to make a version
of your pronunciation,
573
00:28:53,091 --> 00:28:54,550
then please feel free.
574
00:28:54,550 --> 00:28:57,750
Thank you. (laughing)
(applause)
575
00:29:09,131 --> 00:29:10,673
So, I thought I was--
576
00:29:10,673 --> 00:29:13,386
I had the laziest present
but then Andy beat me to it.
577
00:29:13,386 --> 00:29:14,884
(laughing)
578
00:29:15,193 --> 00:29:18,355
So, because I literally made this present
an hour ago.
579
00:29:18,648 --> 00:29:20,561
Some of you might know VizQuery
580
00:29:20,561 --> 00:29:24,640
which is a tool I made
to visually query "Wiki-dah-ta,"
581
00:29:25,218 --> 00:29:28,896
and now I saw this tweet from Maarten
just an hour ago
582
00:29:28,896 --> 00:29:31,428
saying, "Hey, there's a preview
of the Commons Query Service"
583
00:29:31,428 --> 00:29:33,564
so I thought what would happen
584
00:29:33,564 --> 00:29:36,379
if I would just change my SPARQL
end point,
585
00:29:36,379 --> 00:29:40,315
and my tool to the beta
Commons SPARQL end point,
586
00:29:40,761 --> 00:29:43,602
and just add it to my tool,
of course.
587
00:29:43,810 --> 00:29:45,950
Now we need to wait for the wifi.
588
00:29:46,562 --> 00:29:48,752
I should have made a video
but of course,
589
00:29:48,752 --> 00:29:50,993
given that I just had an hour,
here we are.
590
00:29:50,993 --> 00:29:53,430
So, for those of you
who don't know VizQuery,
591
00:29:53,622 --> 00:29:57,041
it allows you to do things like say,
"depicts"
592
00:29:57,041 --> 00:29:59,323
and say, it "depicts a cat,"
593
00:30:00,091 --> 00:30:03,335
and so what you get
are all the Wikidata items
594
00:30:03,335 --> 00:30:05,199
that depict a cat with pictures.
595
00:30:05,199 --> 00:30:08,098
However, what you can do now
is you go all the way down
596
00:30:08,098 --> 00:30:10,159
there's a link saying
597
00:30:10,159 --> 00:30:12,829
use the Wikimedia Commons SPARQL
endpoint experimental.
598
00:30:13,175 --> 00:30:17,451
And now when I say "depicts a cat,"
599
00:30:19,578 --> 00:30:21,114
you will actually get
600
00:30:22,650 --> 00:30:25,346
- Commons images of cats.
- (audience) Woo!
601
00:30:25,346 --> 00:30:26,964
(applause)
602
00:30:30,982 --> 00:30:35,876
So, let's say you want a cat
that actually shows its whiskers.
603
00:30:37,026 --> 00:30:38,885
Now we're going to get...
604
00:30:38,885 --> 00:30:40,001
That's it.
605
00:30:41,003 --> 00:30:42,545
So, well. Thank you.
606
00:30:42,545 --> 00:30:44,378
(laughing)
(applause)
607
00:30:51,935 --> 00:30:53,671
An hour ago well--
(snickers)
608
00:30:53,671 --> 00:30:55,238
That was ten minutes ago.
(laughing)
609
00:30:56,286 --> 00:31:00,644
So, sometime ago I did this tour,
The Wikimedia hackathon in Prague
610
00:31:00,644 --> 00:31:01,755
called inteGraality
611
00:31:01,755 --> 00:31:03,777
making dashboards
of property coverage,
612
00:31:03,777 --> 00:31:08,006
and I introduced to you
the service pack update 2019.
613
00:31:08,006 --> 00:31:09,036
(laughing)
614
00:31:09,467 --> 00:31:11,799
So this is InteGraality,
so you haven't seen it yet.
615
00:31:12,252 --> 00:31:15,325
It makes things like this--
ah it's cute--
616
00:31:15,360 --> 00:31:17,971
or paintings
and their columns of properties
617
00:31:17,971 --> 00:31:21,429
and lines or different groupings.
618
00:31:21,850 --> 00:31:24,536
So, how to slice and dice the data,
619
00:31:24,536 --> 00:31:28,006
and I bring you a couple of improvements
620
00:31:28,167 --> 00:31:30,583
that are going to be live-demoed.
621
00:31:32,110 --> 00:31:34,964
So some people wanted
to be able to query for qualifiers
622
00:31:34,964 --> 00:31:36,736
because some properties
are not top level.
623
00:31:36,736 --> 00:31:38,267
So if we do this...
624
00:31:39,313 --> 00:31:40,587
(laughing)
625
00:31:40,691 --> 00:31:41,882
and...
626
00:31:42,243 --> 00:31:44,508
also some people wanted to display images,
627
00:31:44,508 --> 00:31:46,864
which is I guess is not
the greatest display.
628
00:31:47,819 --> 00:31:49,237
Alright. Loading.
629
00:31:49,237 --> 00:31:50,570
It's supposed to be fast.
630
00:31:50,570 --> 00:31:53,046
Supposed to be fast (laughs)
631
00:31:53,243 --> 00:31:55,426
I know it works
because I already did it.
632
00:31:55,788 --> 00:31:57,242
Yes, updated page.
633
00:31:58,041 --> 00:31:59,272
Yep and that works.
634
00:31:59,272 --> 00:32:02,218
Now the street number.
That worked. (applause)
635
00:32:04,211 --> 00:32:06,286
Pictures, maybe
you're going to make it.
636
00:32:07,700 --> 00:32:09,510
Ah. Nah.
637
00:32:09,910 --> 00:32:12,454
Well, everything [inaudible].
(laughing)
638
00:32:12,538 --> 00:32:13,962
really works.
(laughing)
639
00:32:14,743 --> 00:32:16,438
Yep. Yep. That worked.
640
00:32:17,031 --> 00:32:19,859
Yeah, also works with images.
(audience) Woo!
641
00:32:19,859 --> 00:32:21,116
(applause)
642
00:32:21,116 --> 00:32:23,646
So there were two picture requests
but they were not the worst
643
00:32:23,646 --> 00:32:25,234
and this one was literally done--
644
00:32:25,234 --> 00:32:26,725
Oh what could be that link here?
645
00:32:26,725 --> 00:32:27,750
I wonder.
646
00:32:28,102 --> 00:32:30,901
Okay. Not this one is going
to be too big.
647
00:32:31,675 --> 00:32:34,897
Let's go for yeah--
this is going to be fine.
648
00:32:36,049 --> 00:32:37,916
Yeah, the reason why I spend
my entire time
649
00:32:37,916 --> 00:32:40,555
at the conference doing this
is because I spend the last few weeks
650
00:32:40,555 --> 00:32:43,439
writing tests for all the code
that I wrote in Prague,
651
00:32:43,439 --> 00:32:46,126
and it's like--
Oh, so what could these links be?
652
00:32:48,514 --> 00:32:51,026
Yeah, these are the items
that have the property,
653
00:32:51,026 --> 00:32:52,464
and if you go the other one
654
00:32:52,464 --> 00:32:54,384
that are the items
that don't have the property,
655
00:32:54,384 --> 00:32:55,873
so you can actually make this--
656
00:32:57,142 --> 00:32:58,224
dashboard completely blue,
657
00:32:58,224 --> 00:32:59,954
if you spend enough time.
658
00:32:59,954 --> 00:33:01,042
(laughing)
659
00:33:01,428 --> 00:33:03,886
Yeah. That's the service pack update.
(applause)
660
00:33:18,855 --> 00:33:20,789
Oh, we're at the end of the slides.
661
00:33:21,058 --> 00:33:23,432
Now we're taking the people
who didn't give me slides.
662
00:33:23,432 --> 00:33:25,095
(laughing)
663
00:33:30,284 --> 00:33:31,366
Alright.
664
00:33:31,366 --> 00:33:32,604
Who would that be?
665
00:33:32,604 --> 00:33:34,051
I know one.
(laughing)
666
00:33:34,244 --> 00:33:35,383
Are there other people?
667
00:33:35,383 --> 00:33:37,791
(presenter) I added something
at the back--
668
00:33:37,791 --> 00:33:39,528
Uh huh. Two. okay.
669
00:33:39,528 --> 00:33:41,562
Alright. Amir, you go.
670
00:33:43,631 --> 00:33:47,043
Hello. Sorry for a late minute
presentation.
671
00:33:47,043 --> 00:33:49,472
One reason is that
the dashboard was broken
672
00:33:49,472 --> 00:33:51,239
but we were able to fix it.
673
00:33:51,239 --> 00:33:53,907
So a lot of us use Wikidata,
674
00:33:53,907 --> 00:33:56,210
and you see sometimes
it's a little bit slow
675
00:33:56,210 --> 00:33:57,805
when you want to load a page
676
00:33:57,805 --> 00:34:00,772
especially when the item
is very, very big.
677
00:34:00,772 --> 00:34:02,615
So, in the last month--
678
00:34:03,175 --> 00:34:04,992
several people
at Wikimedia Deutschland
679
00:34:05,612 --> 00:34:09,012
like Rosalie, Jakob and me
started working on it,
680
00:34:09,012 --> 00:34:11,092
and improved the performance of Wikidata.
681
00:34:11,820 --> 00:34:14,360
So now we have something to show
it to you.
682
00:34:14,478 --> 00:34:19,737
So I will go to www.wiki [inaudible]
683
00:34:19,737 --> 00:34:23,727
Where is the slash--German keyboard...
(laughing)
684
00:34:26,033 --> 00:34:27,253
Ah Shift + 7.
685
00:34:30,460 --> 00:34:32,047
Ah yeah.
686
00:34:32,830 --> 00:34:33,929
And...
687
00:34:35,853 --> 00:34:36,892
you go to--
688
00:34:36,892 --> 00:34:38,953
so this is called a--
689
00:34:42,621 --> 00:34:43,962
speed index.
690
00:34:43,962 --> 00:34:46,967
This is a speed index
of item of Berlin,
691
00:34:47,162 --> 00:34:48,766
and you see in the past 40 months
692
00:34:48,766 --> 00:34:50,117
it went from 90---
693
00:34:50,235 --> 00:34:54,782
which is defined as--
let me read it out loud.
694
00:34:54,908 --> 00:34:57,610
The speed index is the average time
of a visible part
695
00:34:57,610 --> 00:34:59,785
of a page or display.
696
00:35:00,204 --> 00:35:01,747
It's express in miliseconds
697
00:35:01,747 --> 00:35:04,115
and depends on the site of--
698
00:35:04,115 --> 00:35:05,747
So it used to be around
699
00:35:05,747 --> 00:35:07,321
1 second for item of Berlin.
700
00:35:07,321 --> 00:35:10,383
Now it's around 800 milliseconds,
701
00:35:10,383 --> 00:35:12,425
and this happens not just
on item of Berlin,
702
00:35:12,425 --> 00:35:17,143
but all items and not just on all items--
plus all images and comments.
703
00:35:17,293 --> 00:35:18,395
All of them got better
704
00:35:18,395 --> 00:35:20,214
for 200 miliseconds.
705
00:35:20,518 --> 00:35:21,939
(whistles)
(auidence) Woo hoo!
706
00:35:21,939 --> 00:35:23,466
(applause)
707
00:35:38,946 --> 00:35:40,085
Hi everyone.
708
00:35:41,671 --> 00:35:46,282
So you may remember
from a previous presentation the Hub,
709
00:35:46,282 --> 00:35:50,660
which is a tool to browse the web
with URLs
710
00:35:50,660 --> 00:35:52,529
going through Wikidata
as the hub.
711
00:35:52,529 --> 00:35:56,586
So you could do things like
going from [inaudible] identifier
712
00:35:56,586 --> 00:35:59,809
to some other identifier.
713
00:35:59,809 --> 00:36:03,860
I don't remember what P9,
1,9,3,8 is... (laughing)
714
00:36:03,860 --> 00:36:06,018
but, yeah... Gutenberg. (laughs)
715
00:36:06,311 --> 00:36:08,056
So... (laughing)
716
00:36:08,056 --> 00:36:10,837
So yeah if you know
those identifiers,
717
00:36:10,837 --> 00:36:12,797
you can go from somewhere
to somewhere else,
718
00:36:12,797 --> 00:36:14,591
and, well, do different things.
719
00:36:14,591 --> 00:36:17,279
Go from like can resolve...
720
00:36:17,839 --> 00:36:21,064
Twitter username on Wikidata
721
00:36:21,064 --> 00:36:23,690
and get redirected to the closest
to Wikipedia article.
722
00:36:23,960 --> 00:36:25,564
But not that--
723
00:36:25,564 --> 00:36:29,038
not that many people use
URLs to browse the web
724
00:36:29,038 --> 00:36:31,253
so I thought if people don't
come to the tools
725
00:36:31,253 --> 00:36:32,836
the tools come to them,
726
00:36:32,836 --> 00:36:36,310
and so I did a little script,
727
00:36:36,411 --> 00:36:39,294
and that you will find there at--
728
00:36:41,268 --> 00:36:42,653
da da dum dum
729
00:36:43,314 --> 00:36:48,294
This on meta
which basically takes the identifiers...
730
00:36:49,196 --> 00:36:51,349
from the Hub
731
00:36:51,517 --> 00:36:54,406
to bring them
to your Wikipedia article.
732
00:36:54,406 --> 00:36:55,870
So, if you add the gadgets
733
00:36:55,870 --> 00:36:59,314
you will have a page
that will,
734
00:36:59,314 --> 00:37:02,157
instead of having just those few
things on the side bar,
735
00:37:02,589 --> 00:37:05,231
because it's not enough
to browse the web.
736
00:37:05,926 --> 00:37:09,984
You will have a collection
of... (laughing)
737
00:37:10,128 --> 00:37:12,112
additional links (laughs)
738
00:37:12,251 --> 00:37:16,538
to all over the web,
and so here you will find, for example--
739
00:37:17,060 --> 00:37:18,103
(grunts)
740
00:37:18,103 --> 00:37:20,074
So this is the page for Berlin
741
00:37:20,074 --> 00:37:21,700
and you will have, for example,
742
00:37:21,700 --> 00:37:23,907
Berlin on an open street map,
743
00:37:24,392 --> 00:37:26,729
Berlin on Quora,
744
00:37:27,263 --> 00:37:32,140
Berlin on Swedish Anbytarforum.
(laughing)
745
00:37:32,854 --> 00:37:36,932
Anything and so all those
convenience links
746
00:37:36,932 --> 00:37:39,757
added to every page
that can be resolved
747
00:37:39,757 --> 00:37:41,578
to a Wikidata identifier.
748
00:37:41,972 --> 00:37:43,096
Thank you.
749
00:37:43,096 --> 00:37:44,611
(applause)
750
00:37:50,075 --> 00:37:53,924
So I didn't understand
at the beginning quite much
751
00:37:54,196 --> 00:37:56,303
the format of this thing.
(laughing)
752
00:37:56,303 --> 00:37:59,266
So I just smuggled inside
753
00:38:01,387 --> 00:38:03,711
And I will try to improvise a lot.
754
00:38:08,971 --> 00:38:10,316
So...
755
00:38:17,591 --> 00:38:18,591
No.
756
00:38:21,816 --> 00:38:23,035
(audience) It's loading.
757
00:38:23,035 --> 00:38:24,274
It's loading? Okay.
758
00:38:24,274 --> 00:38:26,256
So... oh yeah it's maybe only a bit slow.
759
00:38:26,256 --> 00:38:31,749
So, I mean there is a lot of data
in Wikipedia...
760
00:38:33,229 --> 00:38:35,088
so a lot of text which contains
761
00:38:35,088 --> 00:38:38,486
in fact information
that you could add to Wikidata,
762
00:38:38,486 --> 00:38:41,086
but it's sometimes difficult to find,
763
00:38:41,086 --> 00:38:45,984
or difficult to import
so what we did is
764
00:38:45,984 --> 00:38:50,598
we used the latest machine
learning algorithm
765
00:38:51,118 --> 00:38:53,509
to given a class, for example,
766
00:38:53,509 --> 00:38:54,890
newspapers,
767
00:38:54,890 --> 00:38:57,554
check which are the newspapers--
768
00:38:58,420 --> 00:39:01,171
sorry the most used properties
from newspapers
769
00:39:01,171 --> 00:39:02,889
like the owner, the publication date,
770
00:39:02,889 --> 00:39:04,107
the language,
771
00:39:04,107 --> 00:39:06,001
and we are going to the corresponding--
772
00:39:06,001 --> 00:39:08,345
if the statement is missing
773
00:39:08,345 --> 00:39:12,215
in the Wikidata item,
774
00:39:12,215 --> 00:39:13,943
we are going to the Wikipedia page
775
00:39:13,943 --> 00:39:18,759
and we're searching automatically
for this missing statement,
776
00:39:19,066 --> 00:39:21,711
and we are proposing to the user
a new fact.
777
00:39:21,711 --> 00:39:23,955
So the user has just to say
yes or no
778
00:39:23,955 --> 00:39:27,769
to this new fact,
and import it to WIkidata.
779
00:39:27,769 --> 00:39:31,390
So, unfortunately,
this web page is too big
780
00:39:31,879 --> 00:39:34,856
to load or the internet connection
is too slow.
781
00:39:34,856 --> 00:39:39,954
So I'm sorry for that
but we will make a tweet soon,
782
00:39:40,444 --> 00:39:41,922
and launch this tool,
783
00:39:41,922 --> 00:39:45,670
and I think it would be a very good tool
to very quickly add
784
00:39:45,670 --> 00:39:48,411
a lot of statements in WIkidata
about entities
785
00:39:48,589 --> 00:39:51,109
that you are not even aware of.
786
00:39:51,350 --> 00:39:53,443
Okay. Thank you very much.
I'm sorry for...
787
00:39:53,443 --> 00:39:55,254
(applause)
788
00:39:59,975 --> 00:40:02,188
How do I get back to the page?
789
00:40:04,264 --> 00:40:08,352
Yes. So one thing that I did
earlier this year to--
790
00:40:08,352 --> 00:40:11,964
and it's called the Revamp
of Wiki Loves Monuments in Brazil
791
00:40:11,964 --> 00:40:14,306
which was the most successful
Wiki Loves Monuments
792
00:40:14,306 --> 00:40:16,941
that's happened in Brazil
[inaudible]
793
00:40:17,260 --> 00:40:18,908
is create this little box here.
794
00:40:19,222 --> 00:40:21,481
So this is pulling information
from Wikidata.
795
00:40:21,481 --> 00:40:23,311
It's replacing the old style
monument IDs
796
00:40:23,311 --> 00:40:26,354
on Commons which is [inaudible].
797
00:40:26,935 --> 00:40:29,872
So it pulls everything from Wikidata,
multilingual of course.
798
00:40:30,156 --> 00:40:33,087
You need to define the Q id
in this case,
799
00:40:33,344 --> 00:40:36,445
but one thing that changed today,
thank you very much
800
00:40:36,445 --> 00:40:38,537
to the Structured Data on Commons team,
801
00:40:38,537 --> 00:40:40,237
is that they're been able
to Lua access
802
00:40:40,237 --> 00:40:42,021
to Structured Data on Commons.
803
00:40:42,021 --> 00:40:43,159
(audience) Yeah!
(applause)
804
00:40:43,159 --> 00:40:44,732
It's fantastic.
805
00:40:47,393 --> 00:40:49,123
So now what you can do
is you can say,
806
00:40:49,123 --> 00:40:51,727
this picture of telescopes--
sorry I like telescopes--
807
00:40:51,832 --> 00:40:55,596
and has Mark II Telescope
and Lovell Telescope in the U.K.,
808
00:40:55,955 --> 00:40:58,138
and if you go to the file information,
809
00:40:58,775 --> 00:41:00,519
and edit the page,
810
00:41:02,048 --> 00:41:04,849
you will see that that--
811
00:41:04,849 --> 00:41:07,526
sorry you probably can't see it so easily.
812
00:41:09,072 --> 00:41:12,211
You just need to do Monument ID/SDC
813
00:41:12,575 --> 00:41:14,732
and you get that information
automatically
814
00:41:14,732 --> 00:41:16,545
through the Structured Data on Commons.
815
00:41:16,545 --> 00:41:18,240
I think this is the first template
816
00:41:18,240 --> 00:41:20,135
that I can actually do this so--
817
00:41:20,376 --> 00:41:22,579
and because it's only
become available today.
818
00:41:22,579 --> 00:41:24,944
So thank you very much
to Structured Data on Commons.
819
00:41:25,202 --> 00:41:26,814
(applause)
820
00:41:34,415 --> 00:41:37,714
Do we have anyone else
who would like to present something?
821
00:41:39,780 --> 00:41:43,997
If not, then thank you so much
just for awesome presents.
822
00:41:44,376 --> 00:41:46,782
Thank you so much for putting
all the time in to them.
823
00:41:46,782 --> 00:41:48,705
They were really great.
824
00:41:48,931 --> 00:41:51,028
(applause)