I am writing a report on statistics and data analysis. For some experiments I created intentionally some data which is not directly related to real world phenomena. Which term is correct to describe such data: artificial data or synthetic data? Is the data artificially created or synthetically created? Or are both adjectives synonyms?

  • 5
    I would call that "dummy data." – Robusto Sep 16 '15 at 15:00
  • 1
    If you call it synthetic or artificial it conveys the impression (to me) that it is not just random but that you have engineered the data to be realistic. If you just generated random stuff then presumably it is not realistic. Therefore I agree with Robusto and/or Marv Mills. – chasly - supports Monica Sep 16 '15 at 15:23
  • @Robusto Dummy data implies to me that the data has no meaning at all. It just acts as a placeholder. However, my data was created with some properties in mind. Therefore, I think the term "dummy data" does not totally fit. See also wikipedia of definition on dummy data. – Funkwecker Sep 16 '15 at 15:56
  • 2
    YMMV. I use dummy data all the time, and I go to great lengths to make it realistic. – Robusto Sep 16 '15 at 15:58
  • 1
    I agree that 'dummy' is the best word here as it perfectly captures the notion the data is fabricated in some way, closely matches what real data would look like and yet bears no relation to real world data. If you post 'Dummy' as an answer I would upvote it over my own offering. – Marv Mills Sep 16 '15 at 16:01

I would call this "dummy data."

From Wikipedia's entry:

[D]ummy data is benign information that does not contain any useful data, but serves to reserve space where real data is nominally present. Dummy data can be used as a placeholder for both testing and operational purposes.

| improve this answer | |

This is often called 'test data'

Test data is data which has been specifically identified for use in tests, typically of a computer program.


That term is understood to mean data that has been created, synthesised, anonymised from "real" data, or is a copy or version of real data, to be used for the purposes of exercising a computer system in a non-production environment. As suggested in a comment by Robusto, "dummy data" also has this meaning.

| improve this answer | |
  • 3
    I disagree. Test data can be real-world data. For example this data set is collected in real-world and is commonly used for testing computer algorithms. – Funkwecker Sep 16 '15 at 15:46
  • @Julian Actually, I agree, I hadn't considered the case where "real data" is used as test data and I will amend my answer accordingly. The fact that it derives from real world data unchanged, does not invalidate my answer though- for the purposes described it is still test data. – Marv Mills Sep 16 '15 at 15:50
  • Of course the term "test data" can be used for real or artifical data. However, I would like to emphasize that the data is not collected or measured in real-world, but articially (or synthetically) created (with some intention in mind). – Funkwecker Sep 16 '15 at 15:55
  • In that case I think Robusto's comment 'Dummy' is better than 'test', 'artificial' or 'synthetic – Marv Mills Sep 16 '15 at 15:58

You should call this "dummy data" or "test data." The terms artificial data and synthetic data are actually used for data that is constructed to be similar to real data but that manipulated such that the underlying real world items cannot be identified.

For example, the US Census Bureau is now creating synthetic data for its Survey of Income and Program Participation. The data are designed to be provide detailed information on household incomes and participation in federal benefit programs, without compromising the privacy of individual participants.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.