privacy-engineering-tools

Other ways to create synthetic data

Besides R and Python packages, there are other ways synthetic data can be created.

Generative data

Most user interfaces are websites that can generate different data types in a tabular format (e.g., .CSV, .JSON), without the need for input datasets (a.k.a. generative/dummy/fake/sample/test data). Examples of such websites are:

• Mockaroo (free up to 1000 rows)
• Generate Data (max. 500 rows via the website, open source)
• Cobbl (up to 10.000 rows)
• Online data generator (free up to 1000 rows)
• Test data generator (up to 5000 rows)

If you use Qualtrics for sending out a survey, you can also generate test responses in Qualtrics directly, which makes sure the responses correspond to your survey’s data types and flow.

Commercially produced synthetic data

There are many companies that offer services to create synthetic data, often including a utility and privacy report and often using Artificial Intelligence. On some companies’ websites, you can try out their algorithm(s) for free.

Here are some examples, although this is by far not an exhaustive list:

• Clearbox AI offers a free demo of their algorithm for tabular data (csv), up to 10 MB and 3 synthetic data projects per week.
• Syntho is a paid service.
• Mostly AI lets you create 100K rows of synthetic tabular data per day for free.
• The company Synthetized has a Scientific Data Kit: a Python package with a free 30-day trial.
• MDClone is a US-based company that generates health-care data. They provide a free demo on request.

This site is open source. Improve this page.