Residential Location from Purchase Patterns

Requirements Clarification & Assessment

1. Objective:

   • Develop a system to deduce a customer's residential location based on their credit card transaction history to identify potentially fraudulent activities occurring outside this area.

2. Key Considerations:

   • Transaction Time and Frequency: Determine patterns in spending, such as frequent transactions at specific locations during certain times of the day or week.
   • Geographical Boundaries: Define what constitutes "outside of the usual residential area" using geographical data.
   • Data Privacy: Ensure customer data is handled in compliance with privacy regulations.

3. Clarifying Questions:

   • Are there existing models for fraud detection, and how do they currently perform?
   • What is the importance and weight of location in the model's decision-making process?
   • How often is transaction data updated, and how is it stored?

4. Data Requirements:

   • Access to historical transaction data for each customer, including timestamps and merchant locations.
   • Additional demographic data, if available, for cross-verification.

5. Constraints:

   • Online transactions should be distinguished from physical ones to avoid false positives.
   • Account for customers who frequently travel or have multiple residences.