What is an entity?

Here is the basic definition of an entity according to Wikipedia:

“An entity is something that exists as itself, as a subject or as an object, actually or potentially, concretely or abstractly, physically or not.”

But… what is an entity in the context of the semantic web?

An entity can be defined as an information unit and can be nearly everything: a place, a person, a historical event, a fictional character, a book, an abstract concept, and so on. In a website, each entity is a single piece of information connected to other pieces of information. This network of information and relationships is a knowledge graph.

Usually, this content structure is hidden within the text of a web page, and even if it appears crystal-clear to human eyes it’s still quite obscure for search engine crawlers. And here is where entities can make the extra-mile in the semantic web, helping search engines understand properly the content and context of a web page. Web entities are structured data with their own properties, ontologies, and relations, which can be defined with the help of vocabularies like schema.org.

This opens a brand new scenario where search engines’ work is led by meaning and intents not just by keyword.

How does WordLift create entities?

Thanks to natural language processing, WordLift detects main entities in a text and helps content writers turning them into data points, connecting these entities with the linked data through existing datasets such as DBpedia.

Starting from a relevant concept any web writer can give his own definition enriching the web of data with his/her own knowledge. Even though entities are detected by the A.I., a human selection is needed to meaningfully pick the entities which are more relevant in the context of a single article or of an entire website.

Since well-curated entities can become a real search magnet, their choice should be strategically driven by the editorial plan.

Any example?

Even this web page, actually, is an entity detected and created with WordLift, which has been created to define the concept of entity itself – well, we’re aware that it’s kind of a loop, in this very specific case. ?

When crawling this page, search engines will find that this entity:

  1. It’s is the ‘sameAs’ DBpedia‘s entity
  2. It’s related with other relevant entities such as semantic web, DBpedia, Wikipedia, natural language processing, linked data, schema.org, JSON-LD, and WordLift
  3. It belongs to WordLift’s Knowledge Graph.

If you go further diving into the knowledge graph, you will also see more details of each related entity, including WordLift itself.