NCpedia’s Linked Data Experiment

Monday, December 2, 2019

Have you ever visited a web page to read about one topic and ended up clicking a link to read about a related topic? Then, have you clicked another link to read about something else that sparks your interest? Most of us have done this at some point because that’s what the Web is good at – linking related topics to each other. However, the relationships (or links) between the data are not always consistent or structured in a helpful way. That’s where Linked Data comes in.

An abstract photo of connected blue lines symbolizing linked data.

What is Linked Data? 

Linked Data is a set of best practices for publishing structured data on the Web. It allows for large sets of data on the Web to be connected to each other in a way that search engines can easily understand and access the relationship between items. One of the benefits of Linked Data is it helps search engines understand when things are the same. For example, many noted people have a name authority record within the Library of Congress. By linking an NCpedia article about Andy Griffith to his authority record, the search engine now knows that we are referring to the same Andy Griffith and can return more relevant search results.

What did we do with Linked Data?

In early 2018, the Government and Heritage Library began a Linked Data project designed to make it easier for search engines to read and understand the Commemorative Landscape articles within the NCpedia collection. We wanted to see if adding Linked Data to a set of pages made a difference in how search engines ranked these pages. There are a few different forms of Linked Data and the approach we chose is “JSON-LD”: “JSON” refers to the way the specialized data is structured, and “LD” stands for “Linked Data.” We created a JSON-LD “object” for each Commemorative Landscape page.

The JSON-LD helps make connections between the data on the pages and related data in other places on the Internet. For example, the Library of Congress also has a subject heading for “Memorials--North Carolina”. We linked all the Commemorative Landscape articles to this subject heading. Now, when search engines look at these pages, they will connect all the articles to Memorials located in North Carolina. This will help return more relevant search results related to North Carolina Monuments and Memorials.

How will the Commemorative Landscapes pages change?

They won’t! Since our Linked Data is only meant for search engines to read, the new structured data is not visible on the pages. But hopefully, the pages will show up higher in search results when you search for the monuments and related topics. We are currently comparing where the Commemorative Landscape pages showed up in search results before we added the Linked Data to after we added the Linked Data. We will publish the results of our experiment on the State Library’s website in July 2020.

What does Linked Data look like?

If you are interested in seeing what JSON-LD objects look like, you can find all of the objects we created here: https://www.ncpedia.org/sites/default/files/commlandlinkeddata.txt

Below is a selected sample of the data from the Battle of Averasboro Union Soldiers Memorial:

{

"@context":"http://schema.org",

"@type":"CreativeWork", 

"@id":"https://www.ncpedia.org/node/11367",

"name":"Battle of Averasboro Union Soldiers Memorial",

"description":"The monument is a simple design of a gray stone marker on a two-level base.  The front of the monument is decorated with a single star, the insignia of the 20th Corps, and commemorates the Union soldiers killed in the battle of Averasboro.  The back of the monument lists the companies involved in the battle.  The monument was proposed and designed in the early 2000's by Averasboro Battlefield Commission, Inc. ", 
"material":["Granite"], 
"keywords":["Civil War"], 
"genre":"http://id.loc.gov/authorities/subjects/sh91003164.html", 
"geo":  { 
"@type":"GeoCoordinates", 
"latitude":"35.26654", 
"longitude":"-78.67198"

While it may look confusing at first, most of it can be read easily. For instance, the “description” line provides background information about the monument, and the “material” line states what the monument is made out of. You will also see the “genre” provides a URL that directs to the “Memorials--North Carolina” subject heading by the Library of Congress.

Where can I find out more about Linked Data?

If you are interested in learning more about Linked Data, you may check out these references: 
https://www.w3.org/standards/semanticweb/data
https://www.w3.org/wiki/LinkedData
https://www.youtube.com/watch?v=4x_xzT5eF5Q
 

 

Author: 
Erin Holmes, Systems Integration Librarian
Michael Millner, Systems Support Librarian