For teams with an existing dataset
Extend training data from a dataset you already have
Upload a table, inspect what Ground learns and generate additional synthetic rows. Explore the workflow first with a deterministic sample of 3,000 fictional loan applications.
Free sample analysis, up to three tries per account. Review the sample in Ground before starting analysis.
From a source table to an inspected synthetic extension.
- 1
Bring a table or try the sample
Upload your training table, or use the same fictional loan sample linked here. Ground analyzes the source before you choose how to extend it.
- 2
Review the learned patterns
Inspect the source profile, column roles and quality report. Choose your output size and review a preview before generating an export.
- 3
Measure usefulness on a real holdout
Compare a real-data baseline with the augmented training set on untouched, representative real data. Check subgroup behavior and leakage before using the extension.
What to check
The sample includes income, credit score, loan purpose and defaulted. It deliberately includes two inconsistent decision dates to inspect during analysis. It contains no real loan applicants.
What this example can tell you
The starter CSV is a designed simulation, not a real lending dataset. Extending it demonstrates the workflow; it does not validate credit decisions or establish a model-quality improvement. A learned extension is not the same as a rule-generated fixture.