You lend to people with little or no bureau history. I build an explainable scorecard from the data you already hold, such as repayment history, savings, mobile money, rent or utility records.
What you get
A data audit and a target definition that fits your book, even with few or zero defaults
A binned, weight-of-evidence scorecard in Python with points and reason codes
A validation report with Gini, KS, stability and cross-validation
Score bands and cut-off options tied to approval rate and expected loss
Clean, tested code and a handover call with your team
Recent work
At Amani Solutions I built the scorecard for a 490-member SACCO book in Addis Ababa, scoring 374 members at a Gini of 0.75.
Scope
The price covers one scorecard on one dataset. Larger scopes, such as several products or monitoring, are quoted after a short call.
You lend to people with little or no bureau history. I build an explainable scorecard from the data you already hold, such as repayment history, savings, mobile money, rent or utility records.
What you get
A data audit and a target definition that fits your book, even with few or zero defaults
A binned, weight-of-evidence scorecard in Python with points and reason codes
A validation report with Gini, KS, stability and cross-validation
Score bands and cut-off options tied to approval rate and expected loss
Clean, tested code and a handover call with your team
Recent work
At Amani Solutions I built the scorecard for a 490-member SACCO book in Addis Ababa, scoring 374 members at a Gini of 0.75.
Scope
The price covers one scorecard on one dataset. Larger scopes, such as several products or monitoring, are quoted after a short call.