4 Types of Data: Nominal, Ordinal, Discrete, Continuous

Data is all around us. Data takes many different forms, from the number of steps we take each day tracked by our fitness watches to the ratings we give products purchased online to the type of cars people in our neighborhood drive. As someone working with data, whether in business, research, or data science and analytics, it’s crucial to understand the different types of data variables you’ll encounter. Buckle up as we dive into the fascinating world of nominal, discrete, ordinal, and continuous data!

Nominal and ordinal data are generally categorical, whereas continuous and discrete data are quantitative.

What is Nominal Data?  

Nominal data is a type of qualitative (categorical) data used to classify or label information into distinct groups without any inherent order, numerical value, or ranking. 

Its key characteristics are:

• No inherent order 
• Mutually exclusive 
• Qualitative nature 
• Numbers as labels 

The categories cannot be ranked.

An observation or item can only belong to one category.

The common examples of such data include :

Nominal data can be analyzed using methods such as: 

• Central tendency 
• Visualization
• Statistical tests

What is Ordinal Data? 

Ordinal data is a type of categorical data where values follow a logical, natural ranking or order, but the exact numerical difference between each level cannot be defined or measured. 

The key characteristics of ordinal data include: 

• Meaningful order 
• Unequal intervals 
• Categorical base 
• Numeric representation 

The categories have a strict sequence from worst to best or lowest to highest.  

Common examples of ordinal data include: 

Also Read: What Is a Data Mesh Architecture and Why Are Singapore Enterprises Adopting It?

What is Discrete Data?

Let’s start with discrete data, shall we? Discrete data is like the shy, introverted sibling in the data family. It includes values that are distinct and separate, with no in-between. Think of it like counting apples – you can have 1 apple, 2 apples, or 10 apples, but not 1.5 apples. The values of a discrete data variable can be listed out and often arise from counting something.

Examples of data discrete in nature include:

• The number of pets you own
• Shoe sizes
• Pages in a book
• Students in a class

Key characteristics of discrete data:

• Distinct, separate values
• Typically, integers resulting from counting
• Often, a finite number of possible values
• Values cannot be subdivided or broken into finer increments

Must Read: Data Analyst vs. Data Scientist vs. Data Engineer in Singapore: Which Career Offers Better Growth, Salary, and Future Opportunities?

What is Continuous Data?

On the flip side, we have continuous data – the extroverted, life-of-the-party sibling. Continuous data can take on any value within a range, including an infinite number of possible fractional values between whole numbers.

Examples of continuous data abound in the world around us:

• Height and weight measurements
• Temperature readings
• Time durations
• Distances

There’s an endless continuum of possibilities with continuous data, even between two values. A person’s height could be 5 foot 7 inches, or 5 foot 7.1 inches, or 5 foot 7.1052 inches, and so on to infinite precision. The key is that continuous data can be measured to finer and finer increments, depending on the precision of our measurement tools.

Continuous vs. Discrete Data

Now that we’ve met both discrete and continuous data, let’s see how they stack up head-to-head:

Discrete Data:

• Distinct, separate values
• Usually integers
• Result of counting
• Often, a finite number of values

Continuous Data:

• Any value in a range
• Measured to a level of precision
• Infinitely divisible
• Often a measurement like height, weight, time

While discrete data is great for counting and categorizing, continuous data shines when we want to measure and quantify things precisely. The type of data determines the types of visualizations, analyses, and statistical methods most appropriate.

Also Read: Freelance Data Science Consulting in Singapore: Is It a Viable Career Alternative to Full-Time Employment

Nominal vs. Ordinal Data

Beyond the discrete vs continuous distinction, there are two other key data types to know: nominal and ordinal. Nominal data consists of categories without any order or ranking between them. Classic examples include:

• Marital status: Single, Married, Divorced
• Eye color: Blue, Green, Brown
• Zipcodes: 90210, 10001, 60611

With nominal data, the categories have no inherent ordering or hierarchy. Married is not “greater than” or “less than” Divorced.

Ordinal data, on the other hand, has a ranked ordering between categories. The categories themselves are discrete, but they have a clear sequence or ordering.

Examples of ordinal data include:

• Economic status: Low, Medium, High
• Satisfaction rating: Very Unsatisfied, Unsatisfied, Neutral, Satisfied, Very Satisfied
• Frequency scales: Never, Rarely, Sometimes, Often, Always

While we can order ordinal categories, the spacing between them isn’t necessarily equal or measurable. The difference between Low and Medium economic status isn’t necessarily the same as the difference between Medium and High.

Encoding Data Types for Machine Learning

Encoding changes categorical data into numerical representations that machine learning algorithms can process. 

• Nominal data, which has no inherent order, is commonly encoded using one-hot encoding. For example, categories such as “red”, “blue”, and “green” can become separate binary columns. 

• Scikit-learn explains that OneHotEncoder creates a binary column for each category and is useful for many estimators, including support vector machines (SVMs) and linear models.   

• Ordinal data contains meaningful rankings, so ordinal encoding can represent categories with ordered numerical values, such as “low”, “medium”, and “high”. Numerical continuous and discrete data generally already have numerical representatives and may instead require transformation or scaling.  

• For high-cardinality categorical variables, alternatives such as target encoding can reduce the number of generated features. 

Read More: Exploring Key Types of Quantitative Research in Dept

Putting It All Together

We know it might feel overwhelming at first, but understanding these data types is essential if you want to work with data like a pro. Once you know what types of data you’re working with, you’ll be able to:

• Create visualizations that truly showcase your data’s story
• Choose statistical methods that are the perfect fit for your data type
• Sidestep embarrassing data blunders, like trying to calculate the average of phone numbers (we’ve all been there!)
• Talk about your data with clarity and precision

FAQs on Four Types of Data