navedanalytics/african-gig-economy-digital-wallet-risk-analytics

An end-to-end data analytics project analyzing transaction performance, fraud exposure, digital wallet activity, risk patterns, and operational performance across African markets using SQL, Python, and Power BI. # The African Gig Economy & Digital Wallet Risk Dashboard **Author:** Naved Khan **Tools:** Power BI (dashboard) · Excel/SQL/Python (data prep) · DAX (measures) **Type:** Portfolio analytics project — fraud & risk analytics for digital wallets used by gig economy workers across Ghana, Kenya, Nigeria, and South Africa --- ## 📊 Dashboard Preview ## 1. Business Problem Digital wallets are the primary way gig workers (ride-hailing drivers, delivery couriers, domestic workers, freelancers, agricultural day labourers, etc.) get paid across Africa. That convenience comes with exposure: thin KYC files, irregular/high-velocity cash-in/cash-out patterns, and multiple payment rails (Mobile App, USSD, Agent, POS, API/Third-Party) create fraud surface area that's expensive for wallet providers and risky for workers who depend on uninterrupted access to their earnings. **The core question this project answers:** > Where is fraud loss concentrated (market, channel, KYC tier, worker segment), what's driving low transaction completion, and what should the business actually *do* about it? ## 2. Dashboard Structure Three linked pages, filterable by Year/Month: | Page | Focus | |---|---| | **Executive Overview** | Volume, value, fraud loss, and completion trends at a glance | | **Channel & Market Performance** | Channel efficiency, processing speed, transaction-type revenue mix, country-level regulatory tiers | | **Fraud & Risk Intelligence** | KYC-tier exposure, gig-segment risk, velocity vs. fraud correlation | ## 3. Key Insights 1. **Fraud is flagged on ~1 in 2 transactions.** A 50.29% flagged-fraud rate against a 12.64% completion rate signals a funnel problem, not a niche edge case — this is systemic, not isolated. 2. **No single "bad market."** Ghana, Kenya, Nigeria, and South Africa each carry near-identical transaction value (~$6.2–6.3M) and fraud loss (~$3.1–3.2M). Fraud exposure is structural to the product, not a country-specific anomaly — even …