cdn.media.ccc.de/.../wikidatacon2019-1037-eng-All_the_worlds_GLAM_institutions_-_and_how_we_tried_to_find_them_hd.mp4

(chairman) Good afternoon, everybody.
Welcome back.

We will start our last session
in this room before the closing session.

On stage is Alicia Fagerving
from Wikimedia Sweden,

who will tell us more about finding GLAMs
all around the world.

Don't forget
the collaborative note taking--

there is a link to the Etherpads
on the program,

as you know by now, and please help us

all for the documentation
for this presentation.

Thank you very much.

(Alicia) Thank you.

My name is Alicia Fagerving.

I am a developer at Wikimedia Sverige,

the Swedish chapter
of the Wikimedia Foundation.

And today I am going to talk
about the project

that, I know, some of you
have heard about.

It is a project that we have been
working on for quite a while,

and we've been reaching out
to the global Wikimedia community

to keep you updated.

We have received already
some very enthusiastic feedback,

both earlier on and at this conference,

indicating that there is
indeed a lot of interest

in the potential of Wikidata

as the ultimate global database

of cultural heritage institutions.

The project that we are working on
is called FindingGLAMs,

and its goal is, in short,

to put every museum, gallery,
archive and library

on the Wikidata map.

Wikimedia Sweden
is working on this project

in collaboration
with the Wikimedia Foundation

and with UNESCO,

and with financing
from the Swedish Postcode Foundation,

which is an organization that promotes
positive social development.

Originally we thought that we would
end our project in November,

which is why the title of the session,

All the world's GLAM institutions
and how we try to find them,

is in the past tense.

Because we expected that at this point
we would be able take a look back

and sum up our efforts.

Well, it turns out we still have
a lot of work to look forward to

as the project has been extended
until the end of February 2020.

And this is great news as it means
that we will be able

to guide and support more volunteers

who want to bring this project
to their local Wikimedia communities.

And I'll talk about that a bit later on,

but first let's have a look at some facts

to understand why we are even doing this.

How many GLAMs are there on Wikidata?

Almost 130,000 right now.

And that is a lot,
like I would be very happy

if I had 130,000 of anything

but that's obviously not enough.

Now, we don't really know
how many GLAMs there are in the world.

But we have access
to some estimates and service,

for example, IFLA,

the International Federation
of Library Associations and Institutions,

estimates on their website

that there are
about 2.5 million libraries

around the world.

If I remember correctly,
most of these are school libraries.

And on Wikidata

there are about 57,000 items

for different sorts of libraries
around the world.

If we look at museums in the U.S.,

we know that there are
at least 30,000 museums.

That is in the U.S. alone,

while Wikidata only has items
for a couple of thousand,

about 4,000 museums in the U.S.

So this shows
that there is still a lot to do

and a lot of possibilities
to ingest more data.

Obviously it's not only numbers
that are interesting

but where those GLAMs are located.

About 60,000 of GLAMs
on Wikidata have coordinates,

which means we can show them on the map.

And this is an interesting way
of demonstrating

not only the size of the data
but also the geographical bias.

There are a lot of GLAMs
in Western and Central Europe.

You can see that the U.K.,
France, Germany, also Sweden

are well represented.

We can see, there is
quite a lot of dots in the U.S.

and in certain parts of South America.

If we look at Asia,
we see that Japan stands out,

but countries like India,
Russia or Indonesia

have really bad coverage,

despite the fact that they are
very, very highly populated.

And obviously Africa--

I don't even have to mention

that the coverage there
is very, very bad--

very few GLAMs in African countries.

And of course this distribution
is not unique to just GLAMs,

it's typical for other types
of items on Wikidata.

The global North is overrepresented

because that's where
a lot of the volunteers come from,

and that's the areas
they have a lot of interest in.

Now, I haven't mentioned the colors yet

but they are important as well.

We've run the queries
to find the GLAMs in Wikidata twice--

once back in April

and, again, very recently in October.

And the yellow dots
are just the April coverage,

and the purple ones are the new ones
that appeared in the October results.

This shows that a lot of new GLAMs
appeared in that couple of months,

the six months period.

Some of this increase,
not all of it but some of it,

is thanks to our project.

And that's specifically
the U.S. and Africa coverage.

So how has this come about?

A very large part of our work

is working with existing
published datasets of GLAMs

in different countries.

And there are lots of them
in different formats

and published by various
regional or national entities.

For example, in the U.S.

we have the Institute of Museum
and Library Services,

which publishes the results
of its annual Public Library Survey.

It contains data about the country's
public library systems

and public library outlets

that are organized in those systems.

And since this data is produced
by a U.S. government agency,

then it's in public domain and we know
we can use it on Wikidata.

So it was possible for us
to import that data.

Another example of data
that we could import

is data from the Advancing
Library Visibility in Africa project,

which is carried out
by the University of Washington.

And I think this project
is very interesting

because they have quite a similar goal.

They collect data about libraries
in several African countries,

and this includes basic information,

such as the name of the library,
where it's located

and its geographical coordinates.

And the researchers
licensed their data as CC0

to ensure that really it can be reused
as widely as possible

and reach as many people as possible.

So because of that we could add

a couple of hundred
new Wikidata items for libraries

on a continent that had
an extremely poor coverage

of libraries and GLAMs in general.

These are the positive examples

but the truth is that most
of the datasets we found

through our research

could not be imported into Wikidata.

We decided to only work with datasets
that are explicitly licensed

in a way that is compatible with Wikidata,

such as public domain or CC0.

But we believe that
even if a dataset is copyrighted,

it's still valuable,

it's still good
for the Wikimedia community

to know that it exists.

And that's why one of the goals
of the FindingGLAMs project

is collating a master list

of datasets of GLAM institutions

which you can find on Meta,
and if you know of a dataset

from your country or region
that is not there,

you are very welcome to add it there.

Simply knowing that it exists
is very valuable.

For example it can be used as a source
to write Wikipedia articles,

and it provides an entry point
to possibly discuss with the publisher

whether it would be possible to release it
under an open license.

And here is an example.

When researching
GLAM data around the world,

we found that Archives Portal Europe

has a directory
of about 2,000 archival institutions

all around Europe,
including very valuable data

like their coordinates,
addresses and collection sizes.

When we reached out to them

about potentially ingesting
that dataset on Wikidata,

they first said that,

well, it's copyrighted,
it's not open data.

But we told them
about Wikidata and open data,

and about our project,

and they were actually very positive,

and they promised to take steps
to release their dataset

under a CC0 license.

We are having a very positive dialog
around that right now,

and we really hope
that this will become reality very soon.

Now, using official datasets
is only one way

in which we can improve the coverage
of GLAM institutions on Wikidata.

Another one that we are
developing in our project

is making it easier for people,
for new users,

especially users within GLAM institutions,

to add new information to Wikidata.

We would love to see
more GLAM professionals edit Wikidata

and add information
about GLAM institutions.

And many of them
have heard about Wikidata.

Many of them are curious
about how they can use it

in their everyday work.

And editing Wikidata
is a great way to learn that.

But anyone who has tried to introduce

non-Wikimedians to Wikidata

knows that there is
quite a bit of a threshold

before they can start making useful edits,

like even if you don't want to delve
into the technical details

and the data modeling,
there is still a lot to learn

about how the data
in Wikidata is structured,

how to find whatever that interests you,

for example how to find
all libraries in a city.

And, most importantly,

how this particular sort of stuff
normally is modeled

because there are a lot of properties
and lots of good practice

but for a new user
who just created their first blank item,

this is very, very hard and complex,

and it's not clear where to find
this information and whom to ask.

So they often need
personal guidance to learn like

what sort of properties
you add to an item

that represents a museum or a library.

So that's why we are developing software

to make it easier for absolute beginners

in GLAM institutions
to make their first Wikidata edits.

This app is called Monumental.

We developed a first version
that was released earlier this year

and based on the feedback we received.

We are working
on a new improved version

that should be online soon.

I don't think the usable version
is released already.

And the aim of Monumental

is to make searching, viewing
and editing GLAM items easier.

For example the user can see
all the GLAM institutions around them

or in a selected location

without having to write SPARQL queries.

And when it comes to editing,

we think this is especially important:

The user does not have to guess
which properties to use.

The interface displays
a selection of properties

that are normally used with GLAM items,

and the user can see
that even if there are no values

for those properties,

then the user knows
that this information is missing,

and this is what they can add.

And that way they can make
valuable contributions

within this very small limited model,
which reduces confusion

and reduces the fear
of doing something wrong.

And we want to promote this tool

and get as many people
as possible involved.

So we are planning
a global crowdsourcing campaign

targeting primarily GLAM staff.

That is the people who have
the up-to-date relevant knowledge,

hoping that we can make it
as easy as possible for them

to start editing.

And we will be asking people
in the global Wikimedia community

to help out and spread the word
about the campaign

in their local communities.

We imagine that there are
people around the world

who have contact with GLAM institutions

and who have the knowledge
about their local Wikimedia communities

that can help out for example
if translating informational material

or with organizing edit-a-thons
for GLAM professionals.

And you can find
more information about that

on the Finding GLAMs site on Meta.

And we think that what we are doing

is obviously a great goal to pursue

because Wikidata is a powerful platform

which enables and encourages re-use.

So the data that we collect will benefit

not only us in the Wikimedia community

but also everyone who is interested
in where there are

cultural heritage institutions
in the world.

People such as researchers, journalists

will be able to see
and re-use that information.

The data will also be available
to search engines,

which can be especially interesting
for smaller GLAMs

that do not have
a strong presence on the Web.

In Wikidata they can become visible.

And for us in the Wikimedia project

getting more GLAMs on Wikidata,

it will enable us
to better work with them,

for example on Wikipedia.

It would provide material
for example for writing articles

and it can benefit especially
Wikipedias in languages

with smaller communities

which rely more and more
on data from Wikidata

in order to create and enrich articles.

And this will bring
more information about GLAMs

to Wikipedias that have
smaller communities

and not enough resources
to write and maintain those articles.

And, as I said, you can find

more information
about our project in general

and about the campaign
that we are planning

on Meta.

Thank you.

(applause)

(chairman) Thank you very much.

Does somebody have any questions? Yeah.

(Alicia) Yes.

(man 1) Thank you.

Just kind of looking

at the the number
of libraries you've gotten,

there are more libraries
than are galleries.

I'm going to make
the wild assumption that there are.

(Alicia) On Wikidata or in general?

- (man 1) In the world.
- (Alicia) In the world?

Oh yeah. That's also a very...
it's a bit of a weird number from IFLA.

If you got their website, they have
like "a map of coverage of libraries,"

and I do think
that most of those 2.5 million,

they are school libraries.

(man 1) Right.

(Alicia) They do show
like how, what types...

There are significantly more
school libraries than public libraries.

(man 1) But I was thinking

because there are
so many more libraries

than there are, say, art galleries...

Yes.

(man 1) Like why did you
limit the project to...?

So why is it not a libraries project,

for a start, or why is it a GLAM project

and where does GLAM end because--

Oh yeah.

(man 1) So you've got galleries,
archives and museums, and stuff,

but in terms of what we consider GLAM,

like it's a bit more expansive
than that,

so is there kind of...
would there be any ambition

to perhaps, add another letter
to the acronym or something?

Well, we are not not working
with expanding the acronym

like we are officially sticking
to the acronym

because, well, that's the easiest.

We got financing for that,

and this is the easiest way
of explaining this

to non-Wikimedians, honestly.

But we also use the term
cultural heritage institution.

And that is a bit broader.

We use that in communication as well.

For example we ingested some data

about Swedish local cultural
heritage associations

that are technically...

well, they are kind of like
archives/museums/associations

of people working in cultural heritage.

And we decided that yes,
this comes under our umbrella.

[chairman] Can you pass him
the microphone?

(man 2) Hello.

As I come from a community
where we have the technical excuse

but not like this organizing...

Do you have any support
from like your international team?

Would you like materials
or some press releases

or letters to send through some contacts

to engage more local institutions?

They can join us

but we can actually make
all technical stuff ourselves.

Yeah, this is a great question,

and like if you check out
the FindingGLAMs page on Meta,

there is a section with information
for regional coordinators

where the point we are hoping to develop
like an information toolkit

that people can translate
into their languages and use

and something that we would really like
to see from other local communities

is communication
about the datasets they have

and the institutions that can help us out.

(chairman) Anybody else?

(woman) Well, I'm just thinking
because I come from the library world

and of course the IFLA statistics,

they come
from the different IFLA sections

that feed up information to IFLA.

But, anyway, in our respective countries

we are interested
in establishing repositories

about up-to-date data
about libraries and other GLAMs

to end... well,
the issues are of different kinds.

There is gathering data, first of all,

keeping them up-to-date,
normalizing those data

because it goes
to the acronyms and things like that.

So then it's also another issue--

the difference between the institution
as corporate body inside

that occupies
or where the collections are.

And just as an additional information,

you mentioned the Archives Portal Europe,
but Europeana is doing also

work on this,
that identifying their providers.

- Exactly.
- So they have more

than 3,000 data providers,

and they at some point, at the beginning,

they didn't deal with the identification
of the data providers.

But the name that each
of the data providers gave for itself

or that Europeana

associated the provider to the collection
that was given to it

was in different forms.

At some point they found the need
to normalize it,

and so they get one
back to the data providers

asking them for the name and variant name

so that it's disambiguation issues,
identification issues

so it's not a trivial problem.

And I'm bringing
that my experience from France

that we have at least three repositories,
three places

because I saw the link to the French

and I was wondering
where did you get data from.

There is an observatory of public lecture

held by the Ministry of Culture--

I have not yet seen it

There is another repository

held by the Union Catalog of France

that identifies all the libraries
that hold collections with information

extended beyond the identification
of the institution as such

but with opening hours
and the collections that they hold.

And there is another--

(chairman) Quick question.

(woman) Well, I'm just wondering about,

first of all, the best data model

to identify these institutions,

and the best way to communicate

with the communities
that do have those data

and keep them up-to-date

so that there is no gap
between what you have

and where the information routes from.

(Alicia) Thank you.

Data modeling that you mentioned,

the distinction between institution
and physical place,

this is a discussion that we have taken up

in the community discussion on Wikidata.

This is something that is definitely
within the interest of the community.

It is being actively discussed.

(woman) Thank you.

Yes.

(chairman) One last very short question
before we have to end here?

(man 3) Excuse me.

Museums and archives
exist on a continuum--

at the big end you have
the British Museum

and the Hermitage, and so on,

but at the other end

you have a garden shed
with a few objects in

or you have glass cases
in the foyer of a company

with some historical artifacts.

Have you a cut off point?
And if so, where is it?

We don't have so much free available data

that we really need
to have a cut off point.

It's rather the exact opposite.

If we get a nice, beautiful free dataset

that has this sort of
very, very granular distinction,

then we will worry about it,

but it's not a problem
that we have encountered yet practically.

(chairman) Okay, thanks a lot
for your questions

and your reactions and answers.

We have to end here

and invite you to the classroom
next to this one for the final session.

Thank you very much.

(applause)