Linked Data

David Wood architected the first large-scale RDF database (http://mulgara.org), re-architected the Persistent URL service (http://purl.org, http://purlz.org) to support Linked Data, and co-founded the Callimachus Project (http://callimachusproject.org). He is co-chair of the World Wide Web Consortium's RDF Working Group (http://w3.org/2011/rdf-wg/).

Marsha Zaidman is Associate Professor Emerita of Computer Science at the University of Mary Washington, where she served as chair of the Department of Computer Science from 1997 to 2009.

Luke Ruth is a Linked Data developer supporting the Callimachus Project (http://callimachusproject.org).

Michael Hausenblas leads the Linked Data Research Centre in Galway, Ireland. He is the project coordinator of the European Commission FP7 Support Action LOD Around-The-Clock (LATC) and other W3C standardization activities.

foreword by Tim Berners-Lee
preface
acknowledgments
about this book
about the cover illustration
Part 1 The Linked Data Web

1 Introducing Linked Data
1.1 Linked Data defined
1.2 What Linked Data won’t do for you
1.3 Linked Data in action
1.4 The Linked Data principles
1.5 The Linking Open Data project
1.6 Describing data
1.7 RDF: a data model for Linked Data
1.8 Anatomy of a Linked Data application
1.9 Summary
2 RDF: the data model for Linked Data
2.1 The Linked Data principles extend RDF
2.2 The RDF data model
2.3 RDF vocabularies
2.4 RDF formats for Linked Data
2.5 Issues related to web servers and published Linked Data
2.6 File types and web servers
2.7 When you have limited control over Apache
2.8 Linked Data platforms
2.9 Summary
3 Consuming Linked Data
3.1 Thinking like the Web
3.2 How to consume Linked Data
3.3 Tools for finding distributed Linked Data
3.4 Aggregating Linked Data
3.5 Crawling the Linked Data Web and aggregating data
3.6 Summary

Part 2 Taming Linked Data

4 Creating Linked Data with FOAF
4.1 Creating a personal FOAF profile
4.2 Adding more content to a FOAF profile
4.3 Publishing your FOAF profile
4.4 Visualization of a FOAF profile
4.5 Application: linking RDF documents using a custom vocabulary
4.6 Summary
5 SPARQL—querying the Linked Data Web
5.1 An overview of a typical SPARQL query
5.2 Querying flat RDF files with SPARQL
5.3 Querying SPARQL endpoints
5.4 Types of SPARQL queries
5.5 SPARQL result formats (XML, JSON)
5.6 Creating web pages from SPARQL queries
5.7 Summary

Part 3 Linked Data in the wild

6 Enhancing results from search engines
6.1 Enhancing HTML by embedding RDFa
6.2 Embedding RDFa using the GoodRelations vocabulary
6.3 Embedding RDFa using the schema.org vocabulary
6.4 How do you choose between using schema.org or GoodRelations?
6.5 Extracting RDFa from HTML and applying SPARQL 155
6.6 Summary
7 RDF database fundamentals
7.1 Classifying RDF databases
7.2 Transforming spreadsheet data to RDF
7.3 Application: collecting Linked Data in an RDF database
7.4 Summary
8 Datasets
8.1 Description of a Project
8.2 Documenting your datasets using VoID
8.3 Sitemaps
8.4 Linking to other people’s data
8.5 Examples of using owl:sameAs to interlink datasets
8.6 Joining Data Hub
8.7 Requesting outgoing links from DBpedia to your dataset
8.8 Summary

Part 4 Pulling it all together

9 Callimachus: a Linked Data management system
9.1 Getting started with Callimachus
9.2 Creating web pages using RDF classes
9.3 Creating and editing class instances
9.4 Application: creating a web page from multiple data sources
9.5 Summary
10 Publishing Linked Data—a recap
10.1 Preparing your data
10.2 Minting URIs
10.3 Selecting vocabularies
10.4 Customizing vocabulary
10.5 Interlinking your data to other datasets
10.6 Publishing your data
10.7 Summary
11 The evolving Web
11.1 The relationship between Linked Data and the Semantic Web
11.2 What’s coming
11.3 Conclusion

appendix A Development environments
appendix B SPARQL results formats
glossary
index

»A friendly introduction to the use and publication of structured data on the WWW.« Tim Berners-Lee, Director of W3C

»A practical guide for integrating and publishing structured data on the Web.« Christofer Weber, NeoGrid

»Takes a complex academic subject and makes it clear and relevant.« Mike Westaway, AstraZeneca

»Highly recommended to all explorers of the Semantic Web.« Rob Crowther, Author of »Hello! HTML5 & CSS«

Linked Data: Structured data on the Web the book is just what Linked Data the technology has needed. It is a friendly introduction to the use and publication of structured data on the World Wide Web. Linked Data was part of my initial vision for the Web and is an important part of the Web’s future. The Web took off as a web of hyperlinked documents which were exciting to read, but which could not effectively be used as data. And, yes, in fact, much of the Web is data-driven, and the data has been hidden on files inside the server. In slides from my wrap-up talk at the very first WWW conference in 1994, I pointed out that while documents talk about people and things, such as a title deed saying who owns a house, the system was not capturing the data—the actual ownership fact—in a way that could be processed. As the Web evolved, and became more driven by data, there has been frustration that changing, hidden data is not exposed to the reader. Linked Data standards allow you to publish data in a way that can be read by people and processed by machines so that previously hidden flows of data become evident. Linked Data may not be as exciting as a hypertext Web to read, but it is more exciting in terms of making everything work more effectively, from business to scientific research. Machines can read, follow, and combine Linked Data much more effectively than they can perform those actions using other forms of data currently on the Web. The role of machines has previously been subservient to the role of people in the technology used to allow people to communicate. Now machines are beginning to become active participants in the communication. Linked Data allows machines to become more useful partners in our daily lives. Linked Data has come of age in the last couple of years. In the last two years we have seen Google announce its Knowledge Graph and adopt the JSON-LD serialization format for Gmail, and produce a large set of terms for general use at schema.org; IBM announce that the DB2 database will become a Linked Data server; and Facebook expose Linked Data via its Graph API. Other large companies and government organizations have followed suit. We have needed a book like this one to introduce Linked Data development to a new and wider group of programmers. Linked Data will provide you with the questions to ask, even if it doesn’t answer them all. It is a great place to begin your study and kick-start your development. I have known Dave Wood for just about a decade. We met when he started his work with the World Wide Web Consortium. We later worked on a Web research project together. Dave has worked tirelessly to develop Semantic Web and Linked Data frameworks since the late 1990s. As a developer, he is well-placed to show others how it is done. The building blocks of Linked Data are not particularly new. The original proposal for the World Wide Web that I wrote in 1989 for my bosses at CERN included hyperlinks with semantics. The proposal read, in part, “The system we need is like a diagram of circles and arrows, where circles and arrows can stand for anything.” In fact, the Enquire program I had written in 1980 captured the relationships between things in a graph. That was the vision. Now Linked Data is delivering on this vision, by adding meaning that computers can process. As we all know, in the basic hypertext Web, the arrows we ended up with all stood for the same thing: “There is some interesting information over here!” Linked Data extends the “document Web” by allowing arrows to stand for anything we can name with a URI. Hyperlinks gain the semantics they need, and, in the process become much more useful. The Web of hypertext-linked documents is complemented by the very powerful Linked Web of Data. Why linked? Well, think of how the value of a Web page is very much a function of what it links to, as well as the inherent value of the information within the Web page. So it is—in a way even more so—also in the Semantic Web of Linked Data. The data itself is valuable, but the links to other data make it much more so. I believe that the Web should evolve to serve all of us, regardless of our nationality, language, economic motivation, or interests. Linked Data is just one part of that evolution. It is not the end—it is just another part of the beginning. There is still plenty to do, so come join us in building the next generation of the Web! Tim Berners-Lee Director of the World Wide Web Consortium (W3C) 3Com Founders Professor of Engineering, Massachusetts Institute of Technology Professor in the Electronics and Computer Science Department, University of Southampton UK

We love the Web and we love the way it’s evolving from the rather simple web of linked documents of the early 1990s into the framework for the world’s information. Representing data on the Web is an obvious, but slightly harder, next step. We each came to the Web in our own ways but came to Linked Data nearly together. David found the Web as a programmer and later as an entrepreneur, Marsha as an educator, and Luke as a student. Marsha and David are old enough to have started computing with punch cards and paper tape. The Web was a very welcome degree of abstraction from ones and zeros. David was introduced to the Web at Digital Equipment Corporation’s fabled Western Research Lab in California in 1993. It was an eye-opener. One of the first large websites showed photos of thousands of pieces of artwork held by the Vatican. Another showed a list of projects that Digital researchers were working on and linked to each of their own individual web servers for detailed documents. David was hooked. Tellingly, it was the project website that he found most interesting. If only you could link into databases and spreadsheets the way you could link to documents. Marsha also found the Web in the early days, when Gopher was the primary search tool and Web browsers worked in a terminal, and she kept up to date with its rapid changes in order to teach new generations of computer scientists. Her career has lasted long enough for her to see the incredible changes wrought by the invention of spreadsheets and databases on decision making, and this fostered an interest in moving data to the Web. Marsha gave David the chance to teach at the University of Mary Washington just as the Linking Open Data project was starting. Luke took the first class offered to U.S. undergraduates on Linked Data in 2011, followed by an independent study and an internship, all with David. He was eventually hired by David to work on Linked Data projects. Luke and David contribute to the Callimachus Project, an open source Linked Data platform described in this book. We’ve used it to build applications for a variety of organizations, from U.S. government agencies and pharmaceutical companies to publishers and health-care companies. Each of those projects is based on the creation, manipulation, and use of Linked Data. We decided to write a Linked Data book for Web developers because there simply wasn’t one. We all had to learn Linked Data from the specifications or by readying academic papers. There are some other books on Linked Data (David edited two of them), but none are aimed specifically at developers. We thought that our combination of real-world development experience and experience teaching technology would result in a useful book. We hope you agree. It’s our privilege to work with a loosely affiliated international group of people working to bring data to the Web. We hope that you’ll read this book and then join us. We can’t wait to see what the Web will become next.