Part of Dataset Foundry — synthetic, documented, citable datasets across domains
Finance
Log in
The tool

Forge a Micro-Credit & Susu Scoring dataset

Choose how many loan records you want and a seed. The same two numbers will always produce the same dataset — write the seed down and cite it in your methodology chapter.

Capped at 20,000 — the API generates synchronously (see ROADMAP.md Phase 1 for the planned async path beyond that).

Optional — leave both blank for the default (365 days from 2024-01-01). Loan disbursement dates are spread across this range; end must be after start.

Two-stage additionally computes loss given default for defaulted loans — the group_guarantee, loan_to_savings_ratio, and loss_given_default_ghs columns only exist in this mode. Single-stage produces the default-probability columns only.

Advanced: customize the model

Same directional relationships described in the methodology — strengthen, weaken, or rebalance them here without touching any code. Defaults are illustrative, anchored against general microfinance/susu literature, not derived from real lending data.

Default probability (frequency)

Loan purpose is table-rated off the loan-purpose base-default-rate anchor (not adjustable here); these loadings multiply on top of it. Missed-contributions is a banded loading — only each band's multiplier is editable here, not its ratio range.

Loss given default (severity) two-stage only

Name field

Off by default — most statistical work has no use for a name column. This is a separate decision from the Country field above: that only decides which name pool a name would be drawn from, not whether this dataset has names in it at all.