cdn.media.ccc.de/.../wikidatacon2019-1034-eng-Wikidata_works_for_Wales_-_Wikidata_and_the_Dictionary_of_Welsh_Biography_hd.mp4

(moderator) Welcome
to this session here this afternoon.

And now we have a talk from Jason Evans,

who is the National Wikimedian of Wales,

and he'll give a presentation
about Wikidata in Welsh Biographies.

The talk will be 20 minutes,
and we have a Q&A session afterwards.

Thank you very much.

Yes. So, I want to talk generally

about how Wikidata is working
in GLAMs in Wales.

And then I really want to focus
on the journey that we've taken

with the Dictionary of Welsh Biography,

and how we've initially
shared data with Wikidata,

and then how we are using this data now
to actually enrich our own services.

So like many people,
like many institutions,

our kind of journey with Wikidata
began as a result of sharing images

to Wikimedia Commons.

We've shared a lot, we've shared
about 17,000 digital images now

to Wikimedia Commons.

We kind of found that Commons
didn't allow us to share really rich data

about the images,

and we had quite good metadata
for a lot of these collections,

and so we turned to Wikidata
to share that data in a structured way

that we felt was going to provide
more value for our users

than just putting it on Commons.

But we've gone on to share data

about lots of different types
of collection.

And I did a count the other day,

and there is now
nearly 50,000 Wikidata items

that are either directly
about something in our collection

or have some kind of National Library
of Wales identifier on them.

So, we've got this growing
sort of corpus of data

that we've been sharing on Wikidata.

And I'll sort of go through
some of the advantages

that we've seen of doing that,

and showing lots
of different types of collections

as I go along.

So, another type of data
that we've been sharing recently--

we did a project around
a manuscript collection that we've got.

And this is a collection
of medieval Welsh manuscripts.

There are 560 volumes in the collection.

And we took the metadata
for that collection,

and we took actually text descriptions,

that were in all sort of
dumped into a metadata field,

and broke it down as much as we could,

using filters and refining the text
to actually pull out data

and make structured data
for these manuscripts.

And you can kind of see
from the visualization

the kind of data
that we now have access to,

through doing that.

So, we can see a lot more
than just information

about the manuscript itself.

We can access data
about the individual works

that are contained within the manuscripts.

We can see the history of the ownership
of the collection, for example.

So, there are lots of different data
that we can access through doing that.

And this collection in particular
really highlights the research potential.

We've been able to do
really interesting queries

and identify some really interesting
patterns in the data,

like pulling together all the scribes

that worked on
the same manuscript, for example.

And some really interesting
data queries that we can do

that we just simply couldn't do before
when this was metadata.

Another project that we've worked on

is sharing data
for every library in Wales.

And we worked with Simon,
one of our volunteers,

and we worked with CILIP,

the Chartered Institute of Library
and Information Professionals,

and they actually shared all the data
that they had for public libraries,

academic libraries, health libraries.

And this now is the front page
of the main website for Welsh Libraries.

And as you can see, we have a Wiki-Map
on that page that has all the libraries.

They are color-coded by the authorities
that run the libraries,

and you can access their websites
directly through the links on that map.

So, you're really starting to see Wikidata
becoming embedded in the GLAM sector

in the work that we're doing in Wales.

And we're now--

as a result of this sort of engaging
with the library community,

we're now getting calls to actually work

with universities and colleges
that train librarians

to provide Wikidata training
for all new librarians,

which is something I am
quite excited about, trying to roll out.

And then on the subject of books,

we are currently, as some of you may know,

trying to share bibliographic data
for the Welsh bibliography.

So, we're starting with the kind
of the easiest 50,000 items

in that collection--

I'm looking at the most
popular publishers, for example.

And so with the help again of Simon
who's sitting at the back over there,

we have been sharing
these records of books.

We've got about 25,000 items now,
either about the editions, the works,

the publishers, the authors--

so we're really excited
about the kind of potential

that we're going to see
from sharing all these data.

And this, I should say as well,
with the books--

it's part of a Welsh
government-funding project.

So they are funding these projects now.

And in the Welsh government strategy
for the Welsh language--

if you look under the kind
of digital aspect,

it explicitly states
that they support Wikipedia,

and they use Wikidata
to spread knowledge of the Welsh language

and to share information
in the Welsh language.

So then we get to the Dictionary
of Welsh Biography.

This is a collection
of around 5,000 biographies

about Wales' most important
and notable people.

And it began as a series
of volumes in the 1930s,

and it's constantly
being added to over time.

There is still an editorial board,

and there are new articles
being produced monthly.

So, for many years, the only way
of accessing these biographies

was to read a book--

which isn't very trendy these days.

So, in 2004,

this dataset or these articles
were made available online

for the first time.

And I managed to use
the Wayback Machine to get a screenshot

of this amazing website from 2004,

when we first made
the stuff available online.

And this was all about reaching
a wider audience

and engaging with more people.

So the website has been
through several iterations

over the last 15 years.

This is the latest version

of the Dictionary
of Welsh Biography website

that was launched last year.

As you can see,
it looks a lot more modern,

it's kind of adaptable for mobile
and all the rest of it.

But, actually, the functionality
of this website

has changed very little in those 15 years.

You can do it a full-text search
of the articles.

You can search by an A to Z
list of authors and subjects.

And you can do a couple
of very simple filters.

So, this is really quite limiting
in terms of the way

you can actually browse and discover
the content in the collection.

So, in 2015, one of the first things
we did with Wikidata

was put very basic data
about the Dictionary of Welsh Biography

into the Mix'n'match tool.

So, we had a team of about 20 volunteers

that on and off were working
on doing the Mix'n'match process

in the library,

and also there's other Wikidata volunteers

who were kindly sort of helping
as we went along.

Once we'd done all the Mix'n'match

we added all the people
who weren't yet in Wikidata,

and then over the last few years,

we gradually have been kind
of going through this text files

that we have on this website

and extracting more and more information,
sometimes quite manually,

sometimes we were able to kind
of semi-automate that process.

So things like places of birth,
dates of birth,

places of education,
the religion of the people,

related family members,

and adding all these extra data
into Wikidata.

And we've also been,
as we were going along,

making sure that all these items
have Welsh language labels.

But not just the items itself,
we'll do like a SPARQL query and say,

"Show me all the occupations

that are attached to people
in the Dictionary of Welsh Biography,

and then we'll make sure
that all those occupations

have Welsh language labels."

And we do the same for places
of birth and death, for example.

So that users can actually
access and use all the information

attached to the Dictionary
of Welsh Biography people in Welsh.

So, this is an example
of the kind of data that we've now got

for the Dictionary of Welsh Biography.

So, there is some
really good information here.

Now, in the articles, for example,
it will say this person

was educated at the University of Glasgow,
the University of Swansea.

But, of course, what you get
through Wikidata

is all the information
associated with that university,

so you can find out the location of it,

you can take the coordinates
and reuse those,

you can see types of occupations,

you can group them by different things,

you can seek connected family members--

so there is a lot of extra information
that we can pull back

from Wikidata as a result of this sharing.

And that allows us to start doing

some really interesting
queries on the data.

So we've been, over time,
kind of trying out different queries.

This one, for example,
is plotting birthplaces--

it's quite dark on the screen.

There is a map of the world under there.

So, this is like the birthplace
of everyone.

You can see Wales is there, basically.

And then the place of occupation
for all the Christian missionaries

in the collection.

So you can see that
they were working all over the world.

It's a really nice way of kind
of demonstrating

the reach and the impact
that those people had.

Now, we can do things
like look at the average lifespan

of people with different occupations
in the Dictionary of Welsh Biography.

Perhaps, not surprisingly,
Welsh princes and kings

did not last very long

and they probably didn't die
a natural death either.

Also, lawyers and translators,
interestingly-- very short lifespan.

And on the other end of the spectrum,
you have genealogists,

hymn writers, and antiquarians,

who were more likely to live
to kind of the age of 70.

So, the idea is that you can do
some really interesting queries.

And the structured nature of Wikidata

means that we're actually
starting to join up

all these collections
that we've shared to Wikidata,

including the Dictionary
of Welsh Biography.

So, we've now made connections

between the Dictionary of Welsh Biography

and the sitters in portraits,
for example, the artists of portraits.

Scribes of manuscripts, authors of books,

booksellers, printers,
all these different kind of people

that exist in different datasets
in our collections

are now being all drawn together

in one kind of interconnected
web of collections.

And it's not just our own collections
that we're joining together.

Wikidata is allowing us
to connect to collections

in other institutions as well.

So, looking at the external identifiers

for the people
in the Dictionary of Welsh Biography,

we can see that over 1,000 of them

also have an Oxford Dictionary
of National Biography entry.

400 or more have archives
at the National Library.

We can connect to portraits
of these people

in the National Portrait Gallery.

We can even see who's been nominated
for a Nobel Prize.

So, Wikidata is actually helping us
to connect with this growing web

of cultural data from around the world,

with kind of Wikidata acting
as this central hub

to pull all these collections together.

So, the next step for us was obviously,

"Well, how can we reuse all this cool data

that we've now got connected
to our collections?"

So when we launched the new edition

of the Welsh Biography website last year,

we started including information
that comes directly from Wikidata,

and I've highlighted here--

this is the kind
of further reading section

at the bottom of an article.

We've pulled in information
about Oxford Dictionary of Biography,

the Wikipedia article.

We linked directly
to the Wikidata item for the person.

And we can start linking
to external IDs such as VIAF.

And, as you saw on the previous slide,

there are loads of other things
that we could add for our users,

and connections that we could give them,

so that they can link out
to external resources.

And we've been trying to maintain
a kind of community engagement element

of all this work, as well.

So, for example, once all this information
is connected with Wikidata,

it's really easy for us
to draw up lists of people

that don't yet have Wikipedia articles
in English and Welsh.

So, we've been running continuous events

and volunteer projects

to actually fill in those gaps--
those missing people on Wikipedia.

So, then, we started thinking

and looking at how cool Histropedia was

and thinking, "How can we use
all this data that we've shared,

all the open images that we've shared

and that other people
have shared on Commons

that could be used to illustrate
the Dictionary of Welsh Biography,

all the Wikipedia articles

that our wonderful volunteers
have been creating?"

"How can we link that all together
in a kind of discovery interface

that really adds value
to the Dictionary of Welsh Biography?"

And Histropedia seemed
like a really good tool

that would be a really good fit for this.

So, I did a kind of a mock-up
of how something like this might look,

and we set about looking for funding
to try and make this a reality.

And the main aim was really--
of doing this would be

to create a really interactive way

for people to search
the Dictionary of Welsh Biography

that isn't just picking a name off a list.

To give people a way of querying the data

without even knowing
that they're writing a query--

you know, it's all very well
if you know SPARQL,

to do interesting queries on this data,

but through having
a series of filters on a timeline,

you can actually do queries
without the need for any SPARQL skills.

We wanted to connect,
we have 5,000 portraits

that we wanted to share to Commons,

and we knew that a lot
of them would connect

and illustrate people
in the Dictionary of Welsh Biography,

so we wanted that
to be part of the project.

And finally, we wanted to make
as much of the code available as possible,

for any product that we made,

so that other people could use
that as a framework

for building their own timelines
on their own collections.

So, last year, we were very lucky
and very grateful

to receive some funding
from the My-D Foundation in Switzerland.

And we got support
from the National Library of Wales.

And we formed a partnership
with Histropedia,

and we set about building this timeline
that we had envisaged

that really pulls together
all the elements of the data

that we'd been sharing,
the articles we've been creating,

and giving people
really added value to the website.

So, we have an early version
of this timeline,

and you guys are amongst
the first people to see this--

it's not live on the website,
but we feel like we're getting close.

And we're really excited
about launching this

on the Dictionary
of Welsh Biography website.

So, you can see here
we have an interactive timeline

we can switch back and forth

and see everyone
in the Dictionary of Welsh Biography.

As you zoom in, you get more entries.

And when we click someone,

you can instantly link

to their Dictionary
of Welsh Biography article.

So, if we scroll down-- there we are.

So, you get the text of the Dictionary
of Welsh Biography article,

and you can also link
to the Wikipedia article

so that you can see the two
and compare the two.

It's funny, sometimes
you have a tiny article on Wikipedia

but a really detailed one
on the Welsh Biography.

Sometimes it's the other way around--

the Welsh Biography article
might have been written in 1940,

and then you have a really good
up-to-date Wikipedia article,

and the two, I feel,
kind of complement each other.

We can also do a search of the timeline.

So if you wanted to look
for a particular person,

you can just search for them

and pick them out
on the timeline straightaway.

That's an interesting portrait,
I've not come across that one.

But there's a lot more to this timeline

than just being able to sort
of swoosh through and select people.

Firstly, we can add layers
of data on top of it.

So the items that you see in black now

are a curated set of events, related,

that were important in Welsh history.

So that we can give some context

to the lives of the people
that you're seeing on the timeline

by putting in important dates,
like the foundation of universities,

or we've got here the composition
of the National Anthem.

And we can click these items,

and we can link
to their Wikipedia articles as well

so that you can see
information about them.

And we can see all this in Welsh.

And there was actually-- when we came
to do the translation of the interface,

there was only really a handful of things
that we needed to translate

because Wikidata has translations
for everything already.

So it saved us a lot of time.

Is it going to load?

Okay.

Well-- there is a Welsh version.

(laughs)

I'm not sure why it didn't want to load.

Just trust me-- it's there.

Cool. And then we get
to the really powerful thing

and the bit I am most excited about
with this timeline--

and that's the filters.

So, you can apply
all sorts of filters to this timeline.

So, if you want to just get rid
of all the pesky men, for example,

we can do that.

The timeline does become quite sparse,

which is something that we are working on
with the Dictionary of Welsh Biography

to try and get more content.

Oh, I've picked all the men
rather than all the women.

Sorry!

There we go.

So you can just get
all the women on the timeline.

But you can do more
than just add a single filter.

You can add multiple filters.

So if I wanted to search
for all the politicians--

that's great, so we get them,
we get all the politicians,

and there are lots of those,
if we zoom in, you can see.

But then, I can filter this further,

and I can look for all the politicians
that were born in Aberystwyth,

which is my hometown.

And we can filter just down to them.

So if you kind of--
for teachers, for example,

being able to filter
to a very defined set of results

is something that could be really useful.

And I'll show you
another interesting query

that I quite like--

so you can search by given name
and family name.

So if we search for everyone
who has the first name John

and the last name Jones,

there are 82 John Jones
in this set of data.

And that gives you some idea
of the disambiguation challenges we have

when we're working on this kind of data.

So, yeah. There's a lot of John Joneses.

There you go, they just keep coming.

Some of them have a middle name
which is helpful,

but others don't.

And again, we could put
the kind of history of Wales events

on top of that as well.

Theoretically, you could add all sorts
of different layers on top of that.

We're really interested in the future
in kind of adding another layer.

So if we've got archives for a person,

you can click on a person
and see a timeline of the documents

that we hold for that person, for example.

There's really kind
of endless possibilities

with this kind of tool.

So, I think those are the main features

of the timeline.

And, yeah. I think we'll leave it there.

And if anyone has any questions,
do feel free to ask.

(moderator) Thank you very much.

(woman 1) I was wondering,
when you said you're working with them

to increase the number
of women biographies,

how exactly are you doing it?

Okay. So, I've actually managed
to kind of worm my way

into the internal committee
for the Dictionary of Welsh Biography,

and also I attend
the Advisory Board Meetings.

So Wikidata makes it very easy
to highlight the issue with gender.

And it's something that they are aware of.

So there's a few things they're doing.

They've now committed to producing--

they'll only publish 50-50 from now on.

So if they publish an article about a man,

they will next publish
an article about a woman,

and they're keeping that balance.

One of the main things that we're hoping

is through first investment
in the new website--

because the old one is kind of tired--

and inclusion of cool new features
like the timeline

that we'll use as a base
for applying for funding--

to get specific funding
for research projects

so that we can commission
more content about women

and just try and encourage people
to write more content

and really highlight the problem
that we are facing.

So that's the kind of thing
that we're doing.

Any other questions?

(woman 2) How are you getting
the information back in--

you said you're putting it
into the online dictionary.

How are you doing that?

So on the website
for the Dictionary of Welsh Biography,

we basically will, periodically,

I think it's once a month
or once every three months,

we're doing a crawl with Wikidata

because obviously, we know
the Q numbers now for everyone,

and we are updating the links
that we have to the Wikipedia articles,

for example, and everything,
kind of periodically

through the Wikidata items.

And they're just being added on
to the website

as part of the data for each article.

(woman 2) Manually or some sort
of automated process?

No. It's kind of automated.

We've got a little script that we'll go
and pull in all the relevant data

and just add it automatically
to all of the articles.

(moderator) Further questions?

Then, again, thank you
very much for your talk.

Thank you.

(applause)