Synthetic Test Data: What It Is & How to Generate It

Every tester has experienced their fair share of frustrations throughout the testing process. Whether it is inefficient or inaccurate tests, inaccessible gateways, or test data that is either incomplete or incorrect, a seamless end-to-end test can often feel like a fantasy. 

While bottlenecks and dependencies are common in a testing environment, there are tools at your disposal that can help alleviate them — and ultimately increase your testing velocity. One of those tools is synthetic test data. 

In this blog, we will discuss what synthetic test data is, the benefits of using synthetic test data, and how to generate it — as well as highlighting BlazeMeter’s game-changing, AI-driven Test Data Pro. 

Experience synthetic test data generation like never before with BlazeMeter’s AI-driven Test Data Pro. Request a custom demo today! 

Request Demo

What is Synthetic Test Data?

Synthetic test data is a fake version of real test data for developing and testing applications. Synthetic test data mimics real data to allow for better data security as well as filling in gaps within a test with information that would otherwise be unavailable.  

Using sensitive information during the testing process can pose significant security risks. Synthetic test data mitigates those security risks by using fake but equally effective data. Not only does it protect sensitive information, but it also greatly expands test coverage to allow for testing against broader sets of data that can often be complex — names, addresses, geolocation, credit card numbers, and beyond. 

Compare use cases for synthetic test data vs. test data masking >> 

Synthetic Test Data Benefits

Real test data can pose several hinderances — it can be inaccurate, incomplete, or not entirely available. You can either wait for better or more complete data, or you can forge ahead with substandard data that make for a substandard test. Either of those options will set you back significantly in valuable time and resources. 

That is why synthetic test data can be so valuable for your testing strategy. Some of the key benefits are: 

• Tailored To Your Needs — Because the data is fake and generated based on your requirements, you can use data tailored specifically to your use cases. 

• Less Is More — You will not need to sift through or collate large quantities of data. The exact amount of data you need is what you will work with. 

• Minimized Risk — The risks involved with handling sensitive information like banking or health records are eliminated. 

• Dependent No More — Working with real data often means waiting on receiving it from another member of your development team. Synthetic test data means you get it when you need it. 

• Money Saved — Synthetic test data can be generated and discarded on a whim. You will not need to spend large amounts of money on data storage. 

Types of Synthetic Test Data

The type of data you will want generated depends greatly on your needs in any given test scenario. And because your test scenarios can vary widely in context, there are a few types of synthetic test data you can generate. 

Sample Data

Sample data is synthetic data at its simplest. It is largely impromptu data created by developers within a testing sprint. Sample data is used primarily to ensure that all the data fields are occupied. The benefit of sample data is that it can be used for a very specific test to produce a desired response or to test a particular feature (credit card number, for example). The downside, however, is that it is poorly suited for large-scale testing because the likelihood of bugs dramatically increases. 

Rule-Based Data

When using test data, there are often very specific parameters around what is required for any given test. Rule-based data is designed to accommodate those parameters. The key distinction with rule-based data is that it is generated more intentionally than sample data. That means the test data generated is directly correlated to data fields — fields like first and last names, addresses, and postal codes. Rule-based data can come in the form of numerical values, reserved words like “NULL,” blank data, long or short data chunks, or data with special characters. 

Anonymized Data

Replacing real data with anonymized data is an excellent way to preserve data security. Anonymizing — or “masking” — real data enables testers to use the “essence” of real data without risk of exposing sensitive information. Retaining the “essence” of real data means replacing real names with fake names or entirely randomized characters. 

Subset Data

Going the route of subset data will help you tailor your synthetic data for your specific needs or use cases. Doing so will create datasets for your unique test environments and simulations while avoiding unnecessary data. Subset data is a great way to address bugs. Unlike anonymized data, however, subset data does not protect the data within a subset — it only minimizes the risk of exposure. 

Large Volume Test Data

Large-scale testing can often require large amounts of data. Manually doing so eats up significant amounts of time, so large volume test data is synthetic test data primarily generated automatically. With this approach, your testing relies less on the specific data itself and more on the sheer volume and velocity of test data being input. Large volume test data is an excellent way to put your application under duress during performance or load testing.

How to Generate Synthetic Data for App Testing

Let’s look at a few use cases for generating data with BlazeMeter for testing your app. We'll cover a number of methods, from using existing datasets to incorporating specialized AI-powered tools.

Use Case #1: You Have Reliable and High-Quality Existing Data

One effective way to generate synthetic test data is by leveraging existing datasets. This approach is particularly useful when you already have access to non-sensitive data that mirrors the structure of the data your application will process. It can be used to generate vast datasets from a limited dataset, without compromising data privacy and confidentiality.

For example, this method can be used when testing an app for inventory management. By loading your existing parts or goods list into BlazeMeter, you can randomly generate a variety of entries, simulating different inventory configurations.

Steps for generating synthetic data from existing datasets:

• Start by identifying relevant, existing data. For example, a parts catalog, a list of company branches, or supplier contact details. The structure of this dataset should align with the type of data your application will process.
• Convert your dataset to CSV format, so you can easily upload and manipulate the data.
• Use the `randFromCSV` function in BlazeMeter to randomize entries from your CSV file.

Following these steps will allow you to create diverse test scenarios without requiring a large, original dataset.

Use Case #2: You Need Standardized, but Sensitive Data

Testing often requires common and standardized fields like names, addresses, birthdates, and even Social Security numbers. For example, for testing an e-commerce app that requires customer information.

However, these are PII and cannot be uploaded into testing systems ‘as-is’. Instead, synthetic standardized data needs to be generated.

BlazeMeter's standard data functions allow creating:

• Social security numbers
• Credit card numbers
• S&P 500 company information
• US-based addresses
• And more

Use Case #3: You Need Specialized Data

In cases where you need unique, specific datasets, AI-powered data generation is required. For example, for testing a healthcare app might require specialized lists of medical conditions, treatment plans, or physician specialties. Additional examples include collateralized debt obligation types or jobs at a golf course.

Generating Synthetic Test Data With BlazeMeter’s Test Data Pro

Artificial intelligence is quickly changing the landscape of the software testing industry. The latest advancements are showing up in a number of testing tools — including synthetic test data generation tools. 

BlazeMeter is at the forefront of synthetic test data generation after the release of Test Data Pro. It is a four-pronged tool designed to simplify the lives of testers and save teams significant resources like time and money. Take a look at the game-changing features Test Data Pro offers and its benefits will be evident: 

• AI-Driven Data Profiler: Find hardcoded data instantly and automatically generate additional similar data from predefined lists. 

• AI-Driven Test Data Creator: Generative AI greatly streamlines test data generation by converting test to test data functions. 

• AI-Assisted Test Data Function Generator: Ditch time-consuming manual coding by instantly generating test data functions with natural language. 

• Chaos Testing: Find system faults you did not even know where there and boost resiliency through AI-powered test data that challenges systems and identify vulnerabilities.

Combining All Three Approaches for Comprehensive Testing

In many cases, a combination of these methods will yield the most robust testing dataset. Using a mix of existing data, standardized fields, and specialized AI-generated data allows you to cover all bases, ensuring the app performs well across different data inputs and use cases.

For example, testing an application for managing patient records:

• Standard Data - Use functions to generate name, address, date of birth, and the Social Security number.
• Existing Data - Load a list of primary care physicians from current data.
• AI-Generated Data - Create randomized diseases and treatments to complete the patient file, ensuring each test profile is unique and representative of real-world complexity.

This combination provides a robust and thorough dataset that simulates real-life application usage, helping you catch issues early and optimize performance across various scenarios.

Bottom Line