Customized Synthetic Data

In this article we cover how DataSDR generates customized Synthetic Data to improve your software development and debugging processes.

You can download a paper we presented at the PharmaSUG 2021 event.

There are several types of data we provide our clients: “Open Data” vs. “Green Data” vs. “Red Data.”
* “Open Data” refers to data freely available from both academic and government agencies.
* “Green Data” is Synthetic Data we generate to closely match the specific requirements defined in the metadata (see our Metadata Repository platform for details).
* “Red Data” is Synthetic Data purposefully created outside of the boundaries and parameters defined in the metadata.

For example: let’s say you need large amounts of Synthetic Data to develop, debug, and demo your EDC (“Electronic Data Capture”) software platform. Here’s how each type of data can be useful to your organization:

First, we define all the data types your software uses into our Metadata Repository (“MDR”). From individual fields up to forms (CRFs) and study definitions (based off Protocols).

In the Develop phase: we generate “Green Data” that matches the specific parameters defined in the MDR. For a particular CRF, we can generate hundreds or even thousands of records that match the exact field structure, field type, and allowed values for each field. For a field “Age”, defined as integer, with a range between 0 and 120, we’ll generate integer values between 0-120 as expected.

In the Debug / Test phase: we generate “Red Data” that explicitly falls outside the expected values. For the “Age” field we’d generate records with string data as opposed to integers. And with integer values above 120. The goal of using “Red Data” is to verify that the system performs the expected checks and validations.

In the Demo phase: your sales team can showcase your software pre-loaded with large amounts of Green Data. Your team will then showcase all areas of your software with realistic, yet totally fake data.

Ready to improve your software?