Royal Library of Belgium

Sven Lieber

Hi everyone,

this week it's all about hands-on! I will talk about a webinar on trying
to fix quality issues in Wikidata by using its ontology. Additionally I
will share a few resources around basics and advanced uses of the query
language SPARQL that is used to query Linked Data, for example from
Wikidata.

::: {.subscription-widget-wrap attrs="{\"url\":\"https://fairdata.substack.com/subscribe?\",\"text\":\"Subscribe\",\"language\":\"en\"}" component-name="SubscribeWidgetToDOM"}
::: {.subscription-widget .show-subscribe}
::: {.preamble}
Thanks for reading FAIR Data Digest! Subscribe for free to receive new
posts and support my work.
:::

::: {.fake-input-wrapper}
::: {.fake-input}
:::

::: {.fake-button}
:::
:::
:::
:::

📅 Wikidata does *not* have a *formal* ontology, yet there are thousands
of classes and properties that are used by the community to describe
data. Last week I attended a very interesting webinar about the ontology
of Wikidata. The webinar covered, h**ow the ontology works, why it is
important and how to find and try to solve data quality issues.**

🖥️ Besides Wikidata, you may store your RDF data in your own (RDF)
database. Then you even have more possibilities to **use some advanced
data management features** like we do in the
[BELTRANS](https://github.com/kbrbe/beltrans-data-integration) project.
On the recent DHBenelux conference, [Daria
Kondakova](https://twitter.com/dareiadareia) and Jakob Kohler were
looking for solutions to describe personas of Greek Mythology in
different contexts. I think that this is a nice use case for
named-graphs. **I will briefly introduce what named graphs are and how
they can be used with SPARQL**. Additionally I will link some resources
to SPARQL basics.

🏢 Last but not least, in my work update I will talk about the
generalization of a Python script to enrich incomplete tabular data in
an automated way. And also how you could do the same, more manually,
with the Excel-like tool **OpenRefine**.

<div>

------------------------------------------------------------------------

</div>

# 📅 Events {#events}

<div>

------------------------------------------------------------------------

</div>

## Webinar about the Ontology of Wikidata

Each entity (page) in Wikidata should have a type or represent a type!
There are usually entities that describe types such as Human
([Q5](https://www.wikidata.org/wiki/Q5)) and entities that represent
*instances* of these types, such as *Paul Otlet*
([Q1868](https://www.wikidata.org/wiki/Q1868)) who is an "instanceOf" a
Human.

Unfortunately there are millions of pages that **miss such basic type
information**. This makes the data **less FAIR and more difficult to
query and analyze**. These kind of issues and potential ways of trying
to fix it was part of the Europeana Data Spaces webinar *The Ontology of
Wikidata, how to interact with it for a better quality* that I have
attended last Tuesday.

In this very hands-on webinar, [Camillo
Pellizzari](https://www.wikidata.org/wiki/Q74862511) presented the
usefulness of Wikidata's ontology (the types mentioned above, but also
the properties used to create subject-property-object triples). For
example, if you query for people with the occupation engineers you only
get a partial result. But if you tweak the SPARQL query a bit to also
include subclasses of Engineers
([Q81096](https://www.wikidata.org/wiki/Q81096)) **you get a more
complete picture**. Subclasses are for example System Engineers
([Q1572123](https://www.wikidata.org/wiki/Q1572123)) or Geological
Engineers ([Q2887387](https://www.wikidata.org/wiki/Q2887387)) that by
the way have subclasses as well. Check yourself:
[Query](https://query.wikidata.org/#SELECT%20%3Fperson%0AWHERE%20%7B%0A%20%20%3Fperson%20wdt%3AP106%20wd%3AQ81096%20.%0A%7D)
to get engineers (45,072 results) vs
[query](https://query.wikidata.org/#SELECT%20%3Fperson%0AWHERE%20%7B%0A%20%20%3Fperson%20wdt%3AP106%2Fwdt%3AP279%2a%20wd%3AQ81096%20.%0A%7D)
to get engineers and subclasses of it (111,316 results).

Camillo presented the amazing tool [Wikidata Graph
Builder](https://angryloki.github.io/wikidata-graph-builder/) that can
be used to visualize content from Wikidata. For example, the following
little visualization shows the engineer class in the type hierarchy (on
the top right you can choose different visualizations).

::: {.captioned-image-container}
<figure>
<a
href="https://substackcdn.com/image/fetch/f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F751c5123-a4a9-4e50-8df9-4b4355cca15c_1893x883.png"
class="image-link is-viewable-img image2" target="_blank"
data-component-name="Image2ToDOM"></a>
<div class="image2-inset">
<img
src="https://substackcdn.com/image/fetch/w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F751c5123-a4a9-4e50-8df9-4b4355cca15c_1893x883.png"
class="sizing-normal"
data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/751c5123-a4a9-4e50-8df9-4b4355cca15c_1893x883.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:679,&quot;width&quot;:1456,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:83750,&quot;alt&quot;:null,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null}"
srcset="https://substackcdn.com/image/fetch/w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F751c5123-a4a9-4e50-8df9-4b4355cca15c_1893x883.png 424w, https://substackcdn.com/image/fetch/w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F751c5123-a4a9-4e50-8df9-4b4355cca15c_1893x883.png 848w, https://substackcdn.com/image/fetch/w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F751c5123-a4a9-4e50-8df9-4b4355cca15c_1893x883.png 1272w, https://substackcdn.com/image/fetch/w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F751c5123-a4a9-4e50-8df9-4b4355cca15c_1893x883.png 1456w"
sizes="100vw" loading="lazy" width="1456" height="679" />
<div class="image-link-expand">
<img
src="data:image/svg+xml;base64,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"
class="lucide lucide-maximize2" />
</div>
</div>
</figure>
:::

But when clicking on the top left on Mode and select Reverse, you will
see the many sub and sub sub classes of engineers (shown below)! Don't
worry, you can zoom in within the tool. [Try it
out](https://angryloki.github.io/wikidata-graph-builder/?item=Q81096&property=P279&mode=reverse&graph_direction=down)
yourself!

::: {.captioned-image-container}
<figure>
<a
href="https://substackcdn.com/image/fetch/f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F202f5ba2-08e2-4167-a35d-cbbac2b70132_1512x287.png"
class="image-link image2" target="_blank"
data-component-name="Image2ToDOM"></a>
<div class="image2-inset">
<img
src="https://substackcdn.com/image/fetch/w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F202f5ba2-08e2-4167-a35d-cbbac2b70132_1512x287.png"
class="sizing-normal"
data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/202f5ba2-08e2-4167-a35d-cbbac2b70132_1512x287.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:276,&quot;width&quot;:1456,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:36404,&quot;alt&quot;:null,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null}"
srcset="https://substackcdn.com/image/fetch/w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F202f5ba2-08e2-4167-a35d-cbbac2b70132_1512x287.png 424w, https://substackcdn.com/image/fetch/w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F202f5ba2-08e2-4167-a35d-cbbac2b70132_1512x287.png 848w, https://substackcdn.com/image/fetch/w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F202f5ba2-08e2-4167-a35d-cbbac2b70132_1512x287.png 1272w, https://substackcdn.com/image/fetch/w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F202f5ba2-08e2-4167-a35d-cbbac2b70132_1512x287.png 1456w"
sizes="100vw" loading="lazy" width="1456" height="276" />
<div>

</div>
</div>
</figure>
:::

Furthermore, Camillo demonstrated how one, based on SPARQL queries **can
identify problematic entity pages**. For example the ones without a
type. And how one can try to manually try to fix it by assigning a
correct type or adding new hierarchy levels in the ontology based on
trustworthy information from other thesauri such as RAMEAU from the
National Library of France.

He stressed that you can only ***try*** to fix it. Sometimes it is very
complex to fully understand what an existing entity represents and how
it relates to other existing entities. **This is also my main takeaway:
humans remain necessary to correctly curate Knowledge Graphs!**

[Read more about the
webinar](https://www.wikidata.org/wiki/User:Ep%C3%ACdosis/Wikidata-intro#Ontology){.button
.primary}

<div>

------------------------------------------------------------------------

</div>

# 🖥️ Coding corner {#coding-corner}

<div>

------------------------------------------------------------------------

</div>

## Advanced Linked Data management with named graphs

With the Resource Description Framework (RDF) you can represent Linked
Data as subject-predicate-object triples. **But what if you have to
represent additional context of those triples?** You can do this with
*Quads, s*tatements with *four* components: subject, predicate, object
and *context*.

You can imagine those named graph context as a statement saying in which
"box" the triple belongs! The following figure illustrates different
named graphs using different colors. It also shows that with the query
language SPARQL you have different ways to query data in those named
graphs.

::: {.captioned-image-container}
<figure>
<a
href="https://substackcdn.com/image/fetch/f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F82161223-69a0-4e84-bb65-877d04800a79_960x373.webp"
class="image-link is-viewable-img image2" target="_blank"
data-component-name="Image2ToDOM"></a>
<div class="image2-inset">
<img
src="https://substackcdn.com/image/fetch/w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F82161223-69a0-4e84-bb65-877d04800a79_960x373.webp"
class="sizing-normal"
data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/82161223-69a0-4e84-bb65-877d04800a79_960x373.webp&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:373,&quot;width&quot;:960,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:null,&quot;alt&quot;:null,&quot;title&quot;:null,&quot;type&quot;:null,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:null}"
srcset="https://substackcdn.com/image/fetch/w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F82161223-69a0-4e84-bb65-877d04800a79_960x373.webp 424w, https://substackcdn.com/image/fetch/w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F82161223-69a0-4e84-bb65-877d04800a79_960x373.webp 848w, https://substackcdn.com/image/fetch/w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F82161223-69a0-4e84-bb65-877d04800a79_960x373.webp 1272w, https://substackcdn.com/image/fetch/w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F82161223-69a0-4e84-bb65-877d04800a79_960x373.webp 1456w"
sizes="100vw" loading="lazy" width="960" height="373" />
<div class="image-link-expand">
<img
src="data:image/svg+xml;base64,PHN2ZyB4bWxucz0iaHR0cDovL3d3dy53My5vcmcvMjAwMC9zdmciIHdpZHRoPSIxNiIgaGVpZ2h0PSIxNiIgdmlld2JveD0iMCAwIDI0IDI0IiBmaWxsPSJub25lIiBzdHJva2U9IiNGRkZGRkYiIHN0cm9rZS13aWR0aD0iMiIgc3Ryb2tlLWxpbmVjYXA9InJvdW5kIiBzdHJva2UtbGluZWpvaW49InJvdW5kIiBjbGFzcz0ibHVjaWRlIGx1Y2lkZS1tYXhpbWl6ZTIgIj48cG9seWxpbmUgcG9pbnRzPSIxNSAzIDIxIDMgMjEgOSI+PC9wb2x5bGluZT48cG9seWxpbmUgcG9pbnRzPSI5IDIxIDMgMjEgMyAxNSI+PC9wb2x5bGluZT48bGluZSB4MT0iMjEiIHgyPSIxNCIgeTE9IjMiIHkyPSIxMCI+PC9saW5lPjxsaW5lIHgxPSIzIiB4Mj0iMTAiIHkxPSIyMSIgeTI9IjE0Ij48L2xpbmU+PC9zdmc+"
class="lucide lucide-maximize2" />
</div>
</div>
</figure>
:::

Named graphs are a very powerful technique to manage RDF triples.
However, different implementations of triple stores and SPARQL engines
treat named graphs differently, **leading to unexpected and
non-interoperable results!**

**How can you use named graphs and how do these different SELECT
statements work?** Check out my new Blog post to find the answer! I
explain and visually illustrate examples inspired by a discussion on the
recent DHBenelux conference and by my work on the BELTRANS project.

[Read the full blog
post](https://sven-lieber.org/en/2023/06/26/rdf-named-graphs/){.button
.primary}

<div>

------------------------------------------------------------------------

</div>

## SPARQL basics at the ISE lecture

While talking about SPARQL, are you interested in learning (or looking
up) some basics? Just last week the slides of the the *Information
Service Engineering (ISE)* lecture on SPARQL from prof. [Harald
Sack](https://sigmoid.social/@lysander07)
([Q30078997](https://www.wikidata.org/wiki/Q30078997)) were made
accessible on
[GoogleDrive](https://drive.google.com/file/d/13gLES1BYGi9E__-tBsuWno6rPffFkR5G/view).

If you are interested in more RDF or simply are more of a video person,
you can also check out a previous edition of the ISE lecture series on
YouTube!

[Watch the 2021 ISE lectures on
YouTube](https://www.youtube.com/@ISEFIZKarlsruhe/playlists?view=50&sort=dd&shelf_id=2){.button
.primary}

<div>

------------------------------------------------------------------------

</div>

# 🏢 Work updates {#work-updates}

<div>

------------------------------------------------------------------------

</div>

Last week I talked about a newly implemented Python script we used to
enrich incomplete tabular data with data from the ISNI database.

I generalized the script such that you can configure also **other
Application Programming Interfaces (APIs)** besides the one from the
ISNI database. The first release of this command line tool is also
available using the following identifier [DOI:
10.5281/zenodo.8075058](https://doi.org/10.5281/zenodo.8075058) (thanks
to the GitHub integration of Zenodo).

However, I have to say that after implementing it, I realized that the
tool OpenRefine can do the same! Thus if you prefer a user interface and
if you are not afraid of adding a few snippets of parsing code to
OpenRefine yourself, you could use it instead of our command line
script.

The added value of our command line script remains: without much
configuration you can call the command line script. Among others, this
allows to **perform the enrichment in an automated fashion** **without
manually interacting** with a user interface.

[How to use OpenRefine to query SRU
APIs](https://www.oclc.org/developer/news/2018/using-oclc-apis-in-open-refine.en.html){.button
.primary}

<div>

------------------------------------------------------------------------

</div>

That's it for this week of the FAIR Data Digest. I hope you found the
content interesting. Don't forget to share or subscribe. See you next
week!

Sven

::: {.subscription-widget-wrap attrs="{\"url\":\"https://fairdata.substack.com/subscribe?\",\"text\":\"Subscribe\",\"language\":\"en\"}" component-name="SubscribeWidgetToDOM"}
::: {.subscription-widget .show-subscribe}
::: {.preamble}
Thanks for reading FAIR Data Digest! Subscribe for free to receive new
posts and support my work.
:::

::: {.fake-input-wrapper}
::: {.fake-input}
:::

::: {.fake-button}
:::
:::
:::
:::


Hi everyone,  this week it’s all about hands-on! I will talk about a webinar on trying to fix quality issues in Wikidata by using its ontology. Additionally I will share a few resources around basics and advanced uses of the query language SPARQL that is used to query Linked Data, for example from Wikidata.

Rogue Scholar Beiträge

FAIR Data Digest #4