Data curation with synthetic data generation

1. A system, comprising:

at least one processor; and

at least one memory including program code which when executed by the at least one processor provides operations comprising:

identifying an identifier field included in a first datatype of a seed data sample associated with a source system, the identifier field storing a first value that enables a differentiation between different instances of the first datatype, wherein the seed data sample comprises data that occurs at the source system;

identifying a relationship field included in the first datatype of the seed data sample, the relationship field storing a second value that defines a relationship between the first datatype of the seed data sample and a second datatype;

generating, based at least on the seed data sample, a first synthetic data sample, the generating includes populating the identifier field of the first synthetic data sample with a first synthetically generated value and the relationship field of the first synthetic data sample with the second value, wherein the first synthetic data sample is generated to supplement a volume and a diversity of the data at the source system; and

sending, to a target system, the first synthetic data sample to enable a performance of a task at the target system;

in response to determining that the first datatype is a parent datatype to a second datatype, propagating, to a second synthetic data sample of the second datatype, a change corresponding to the first synthetically generated value populating the identifier field of the first synthetic data sample.

2. The system of claim 1 , further comprising: in response to determining that the first datatype is a child datatype of a second datatype, propagating, to the identifier field of the first synthetic data sample, a change corresponding to a second synthetically generated value populating the identifier field of a second synthetic data sample of the second datatype.

3. The system of claim 1 , wherein the first synthetic data sample is sent to the target system by at least pushing the first synthetic data sample to an event stream providing a constant flow of data to the target system.

4. The system of claim 1 , wherein the first synthetic data sample is a variation of the seed data sample that is different from the seed data sample but retains a same dependency to other datatypes.

5. The system of claim 1 , wherein the first synthetic data sample is sent to the target system by at least pushing the first synthetic data sample to a raw data store at a data lake platform where the first synthetic data sample undergoes an extract, transform, and load (ETL) process before being ingested by the target system.

6. The system of claim 1 , further comprising: converting the first synthetic data sample from a first format to a second format; and sending, to the target system, the first synthetic data sample in the second format.

7. The system of claim 6 , wherein the first format comprises a data-interchange format, and wherein the second format comprises a column-oriented data storage format.

8. The system of claim 6 , wherein the first format comprises a JavaScript Object Notation (JSON) and/or an Extensible Markup Language (XML), and wherein the second format comprises Parquet.

9. The system of claim 1 , wherein the task includes reporting, visualization, analytics, and/or machine learning.

10. The system of claim 1 , wherein the first synthetic data sample is used to train a machine learning model to perform the task at the target system.

11. The system of claim 1 , wherein the first synthetically generated value is a randomly generated value.

12. A computer-implemented method, comprising:

identifying an identifier field included in a first datatype of a seed data sample associated with a source system, the identifier field storing a first value that enables a differentiation between different instances of the first datatype, wherein the seed data sample comprises data that occurs at the source system;

identifying a relationship field included in the first datatype of the seed data sample, the relationship field storing a second value that defines a relationship between the first datatype of the seed data sample and a second datatype;

generating, based at least on the seed data sample, a first synthetic data sample, the generating includes populating the identifier field of the first synthetic data sample with a first synthetically generated value and the relationship field of the first synthetic data sample with the second value, wherein the first synthetic data sample is generated to supplement a volume and a diversity of the data at the source system; and

sending, to a target system, the first synthetic data sample to enable a performance of a task at the target system;

in response to determining that the first datatype is a parent datatype to a second datatype, propagating, to a second synthetic data sample of the second datatype, a change corresponding to the first synthetically generated value populating the identifier field of the first synthetic data sample.

13. The method of claim 12 , further comprising: in response to determining that the first datatype is a child datatype of a second datatype, propagating, to the identifier field of the first synthetic data sample, a change corresponding to a second synthetically generated value populating the identifier field of a second synthetic data sample of the second datatype.

14. The method of claim 12 , wherein the first synthetic data sample is sent to the target system by at least pushing the first synthetic data sample to an event stream providing a constant flow of data to the target system.

15. The method of claim 12 , wherein the first synthetic data sample is a variation of the seed data sample that is different from the seed data sample but retains a same dependency to other datatypes.

16. The method of claim 12 , wherein the first synthetic data sample is sent to the target system by at least pushing the first synthetic data sample to a raw data store at a data lake platform where the first synthetic data sample undergoes an extract, transform, and load (ETL) process before being ingested by the target system.

17. The method of claim 12 , further comprising: converting the first synthetic data sample from a first format to a second format, the first format comprising a data-interchange format and the second format comprising a column-oriented data storage format; and sending, to the target system, the first synthetic data sample in the second format.

18. A non-transitory computer readable medium storing instructions, which when executed by at least one data processor, result in operations comprising:

identifying an identifier field included in a first datatype of a seed data sample associated with a source system, the identifier field storing a first value that enables a differentiation between different instances of the first datatype, wherein the seed data sample comprises data that occurs at the source system;

identifying a relationship field included in the first datatype of the seed data sample, the relationship field storing a second value that defines a relationship between the first datatype of the seed data sample and a second datatype;

generating, based at least on the seed data sample, a first synthetic data sample, the generating includes populating the identifier field of the first synthetic data sample with a first synthetically generated value and the relationship field of the first synthetic data sample with the second value, wherein the first synthetic data sample is generated to supplement a volume and a diversity of the data at the source system; and

sending, to a target system, the first synthetic data sample to enable a performance of a task at the target system;

in response to determining that the first datatype is a parent datatype to a second datatype, propagating, to a second synthetic data sample of the second datatype, a change corresponding to the first synthetically generated value populating the identifier field of the first synthetic data sample.