Linked Data (also known as the Semantic Web) doesn’t make the headlines like blockchain. Startups aren’t rewriting elevator pitches to bask in its glow as with AI. But don’t mistake the silence for stillness: Linked Data is ready to change business.
Just as today’s web connects documents, which contain information rendered in human-readable natural languages, the web of the future will connect data. Concepts described in machine-readable datasets will link to other data via common references — just as webpages are connected by hyperlinks. People and machines will follow these connections as easily as we browse the web today.
Let’s talk about what you will experience when you visit a Linked Data-driven company of the near future. You will find it curiously difficult to distinguish between “traditional” data workers (analysts, data scientists, etc.) and those in other functional areas who, at other companies, are less reliant on data. The agent of change here is the unambiguous way that Linked Data represents the world. Semantic technology expert Lee Feigenbaum summarizes this idea:
“For people, these technologies use the same language that subject-matter experts in a domain would use to talk about their data. They provide labels and descriptions intended for people, and they’re not obfuscated with irrelevant IDs, codes, or abbreviations. Often, software user interfaces can be driven directly from the human-friendly descriptions of the data in RDF Schema and OWL.”
This is not to say that learning to use Linked Data today is a breeze, but some of the most exciting innovations in the space are closing the gap between the intuitiveness of Linked Data’s structure and the relative difficulty of its exploration. Ontoforce, which has created a wonderfully accessible semantic search platform, is one company contributing to this effort. data.world, for our part, makes the data able to be queried, and each element is assigned a unique name so any two datasets can be queried jointly or merged for analysis.
Returning now to your imagined on-site visit, you will marvel at the volume and variety of data accruing from disparate sources, flowing from team to team, integrating with other data, producing unexpected insights, available to anyone at any time. Speaking at Techonomy 2017, Turner CTO Jeremy Legg described the perils of internal data hoarding and needless silos. The talk struck a chord with attendees because silos are among the most common features of the corporate landscape. Linked Data transcends silos.
Semantic technologies use Uniform Resource Identifiers and shared vocabularies consistently across databases so new data can be added without requiring the costly, complicated changes companies must slog through when adding new data to relational databases, which are purpose-built and notoriously inflexible.
All this will seem oddly familiar, resembling the web we already know in many respects. The data, for example, would be browseable and searchable by humans, crawlable and queryable by machines. Additionally, just like the Web, Linked Data enjoys a remarkable network effect in that each data set added to the network increases the incremental value of every data set in the network.
You will be inspired by the rapid creation and adjustment of models and automated processes in response to real-time data. Much of this agility is fueled by machine learning models being deployed at a far faster pace than can be achieved without the aid of Linked Data. This is because the output of machine learning is tightly correlated with the quality of input data. People who work in this area spend much of their time cleaning and preparing input data, whereas semantically linked data has been “pre-understood” and embedded with knowledge.
The first wave of companies to adopt Linked Data were established players in “semantic stronghold” industries built on massive data sets and deep R&D coffers—primarily pharma and biosciences, aerospace, finance, and Big Tech. Now we’re entering a phase during which practically any company can seize on the opportunity created by vanishing barriers to entry, compounding Big Data network effects, and a dawning realization that the energy devoted to the costliest, slowest phase of data work—preparation—can finally be reallocated to more productive activities like analysis.
