Machine Learning is so fundamental that it could be used to redo almost everything in industry. Understanding the four main data types is the the key you can use to open the door to Machine Learning.
When you work with different data types, you need to gain an understanding of them from a machine learning perspective, especially if you’re working on visualizations and data storytelling. Alina Zhang, our data superhero, shares insights into a machine learning perspective on different data types. Learn from the best to make better decisions about data conversions and encoding.
Alina explains that machine learning generally categorizes data into one of four main data types. These are:
- Numerical data – This can be discrete or continuous data, but it always uses exact numbers that are not ordered in time. It’s also called quantitative data.
- Categorical data – This is data that expresses characteristics, so it is also called the “class label” in a super classification context. Although categorical data can be represented using numbers, the numbers don’t have a mathematical meaning.
- Time series data – This data consists of numbers that were collected across a period of time.
- Text data – This is essentially words, which you might want to turn into numbers as soon as possible.
Alina notes that the type of data you’re handling will impact the type of algorithms you’ll use or questions you’ll ask of the data. Read Alina’s full post to learn more about recognizing these four main data types, as well as about ordinal data, which is a mix of numerical and categorical data.