Ways to represent data

What is data?

Data is a collection of facts (numbers, descriptions, observations) used to record information. It's organized into structures that usually represent entities important to the organization: customers, products, orders. Each entity has one or more attributes: a customer has a name, an address, a phone number.

Depending on how closely it follows a structure, data is classified as structured, semi-structured, or unstructured.

Structured data

Most commonly the schema is tabular: one or more tables with rows (each instance of the entity) and columns (each attribute).

It's usually stored in a database in which multiple tables reference one another using key values, following the relational model. For example, an order stores the key of the customer who placed it instead of repeating all their details.

Semi-structured data

Example: most customers have an email address, but some have several and some have none. A common format is JSON (JavaScript Object Notation):

// Customer 1: phone and email
{
  "firstName": "Joe",
  "lastName": "Jones",
  "contact": [
    { "type": "home",  "number": "555 123-1234" },
    { "type": "email", "address": "joe@litware.com" }
  ]
}

// Customer 2: email only, plus a "unit" field in the address
{
  "firstName": "Samir",
  "lastName": "Nadoy",
  "address": { "streetAddress": "123 Elm Pl.", "unit": "500" },
  "contact": [
    { "type": "email", "address": "samir@northwind.com" }
  ]
}

Both documents describe customers, but the specific fields change. JSON isn't the only option: XML also represents semi-structured data. What matters is the flexibility, not the syntax.

Unstructured data

Not all data has structure. Documents, images, audio, video, and binary files might not have a specific structure: this is unstructured data.

This kind of content is usually stored as binary files (BLOBs) that an application must interpret and render.

Vector data (embeddings)

Organizations increasingly work with vector data, also called embeddings: a numerical representation of content such as text or images.

Comparing the three types

Where data is stored

Organizations store data to record entities (customers, products), events (sales), or other information in documents and images, and then retrieve it for analysis. There are two broad categories of data store:

The following articles in this domain explain file formats, database types, and which Azure service fits each case.

Exam scenarios

Common exam traps