FSP Data Analysis and Anomaly Detection

Purpose and Outcomes

The manual documents an end-to-end process for turning redemption transaction files into beneficiary- and retailer-level datasets.

2

Collect source redemption files, then run the BeneficiaryLevelData and RetailerLevelData VBA processes.
Create an annual MasterData workbook to consolidate records for year-wide analysis and reporting.
Run the anomaly detection VBA and review its outputs in the AnomalyReport sheet.
Investigate flagged cases, archive completed outputs, and document actions taken for auditability.

3

Detect retailer monopoly within a woreda and retailers operating across multiple woredas for coverage review.
Identify merchant mobile numbers shared by multiple retailer names, which may indicate duplicate or linked accounts.
Flag households appearing in multiple woredas and customer mobile numbers shared across multiple households.
Review top concentration risks and unusual patterns to prioritize cases for validation and follow-up.

4

Run the workflow monthly after each redemption cycle to support timely monitoring and investigation.
Schedule quarterly management reviews and an annual audit review to assess trends and control effectiveness.
Keep original source files unchanged and run the VBA assets on a consistent, documented schedule.
Treat each anomaly as an indicator, not proof of fraud; validate findings with field offices and implementation teams.

5

Core VBA assets generate beneficiary data, retailer data, annual MasterData workbooks, and anomaly reports.
Consistent archiving and documentation preserve traceability from source transactions through investigation outcomes.
Future enhancements include automated folder monitoring, scheduled refreshes, Power BI dashboards, and geospatial analysis.
AI risk scoring, predictive detection, and automated investigation summaries could further accelerate oversight.