The process of publishing your data on the Web is multifaceted. To increase the value of your data sources, you want the data to be expressed using standard names (codes), schemas and serialization formats. This makes it easier for the consumers of your data to integrate it into their applications and with other data sources.
Whether your data is tabular or not, you are often using entities (or ids or codes representing them) as values. Such as artists in your collection of artworks.
People often mint IDs for the entities they are describing. This means they create new ID spaces and assigning new IDs to entities.
Instead of generating new IDs, consider using standard keys for entities, such as Cofactor IDs. Using standard keys (Cofactor IDs) in your dataset facilitates the adoption and increases the value of your dataset.
An alternative to IDs are publicly accessible HTTP URIs pointing at some location on the web. These have the following benefits:
Each Cofactor ID is dereferenceable using its corresponding URI.
What are you representing in your data? Is it people? Events? Try to find an existing schema that meets your needs. Consider schema.org. You want your data schema to be widely understood and adopted. Reuse common vocabularies as much as possible.
If schema.org doesn't fulfill your needs, you can create and publish a custom vocabulary. Avoid defining a new vocabulary from scratch, but try to extend an existing vocabulary.
Here’s an example:
tri:reporting_year rdf:type rdf:Property ; rdfs:label "reporting year" ; rdfs:comment "Indicates a year when a TRI Report was submitted to the U.S. EPA's Toxic Release Inventory system." ; rdfs:domain tri:Report ; rdfs:range rdfs:Literal .