Programming the Semantic Web

Chapter 3. Using Semantic Data

So far, you’ve seen how using explicit semantics can make it easier to share your data and extend your existing system as you get new data. In this chapter we’ll show that semantics also makes it easier to develop reusable techniques for querying, exploring, and using data. Capturing semantics in the data itself means that reconfiguring an algorithm to work on a new dataset is often just a matter of changing a few keywords.

We’ll extend the simple triplestore we built in Chapter 2 to support constraint-based querying, simple feed-forward reasoning, and graph searching. In addition, we’ll look at integrating two graphs with different kinds of data but create separate visualizations of the data using tools designed to work with semantic data.

A Simple Query Language

Up to this point, our query methods have looked for patterns within a single triple by setting the subject, predicate, or object to a wildcard. This is useful, but by treating each triple independently, we aren’t able to easily query across a graph of relationships. It is these graph relationships, spanning multiple triples, that we are most interested in working with.

For instance, in Chapter 2 when we wanted to discover which mayors served cities in California, we were forced to run one query to find “cities” (subject) that were “inside” (predicate) “California” (object) and then independently loop through all the cities returned, searching for triples that matched the “city” (subject) and ...