Turning messy financial data into profit-driving dashboards, monitoring systems, and decision-making tools for African businesses.
I build systems that do three things reliably:
- Reveal where money is leaking.
- Show where growth is hiding.
- Help founders and operators make better decisions faster.
A browser-based financial intelligence tool that transforms raw M-Pesa statements into clean analytics and visual insights for business owners and everyday Kenyans.
This is built to solve a real local problem, financial blind spots caused by transaction chaos.
🔗 Live App: https://m-prism.vercel.app/app
What it does:
- Turns raw M-Pesa statements into clean financial dashboards.
- Instantly categorizes income and expenses.
- Helps small business owners track real profitability.
- Works fully in the browser for privacy and speed.
This is not a demo. This is a real product built for Kenyan realities.
Designed and built a full-scale Sales and Finance Business Intelligence system for a growing manufacturing and distribution company.
Scope included:
- Sales performance monitoring.
- Cashflow visibility.
- Customer concentration analysis.
- Finance KPIs for management decision-making.
Built using:
Power BI, SQL, Python, ERP Data Modeling, ETL Pipelines
Company operates in stealth, name withheld by agreement.
Live FinTech analytics engine designed to detect fraud anomalies and model customer LTV in real-time.
- Engineering Feat: Optimized client-side processing of 10,000+ transaction rows using streaming logic.
- Key Feature: Implemented fuzzy duplicate detection to sanitize dirty financial ledgers.
🔗 Live Demo: https://banking-fraud-analytics-suite-gb5j.vercel.app
🔗 Code: https://github.com/Elias-3817/banking-fraud-analytics-suite
Tech: TypeScript, React, Vite, Streaming ETL, Statistical Anomaly Detection
Machine learning system that predicts customer churn and flags high-risk accounts.
Estimated revenue protection of about $128,000 per month in lost customers.
🔗 Code: https://github.com/Elias-3817/Churn-customer-prediction
Tech: Python, XGBoost, Random Forest, Cost-Sensitive Modeling
Interactive dashboard that exposed $1.2M in revenue risk from late deliveries and revealed that 5 percent of customers generated 26 percent of total revenue.
🔗 Code: https://github.com/Elias-3817/Olist-Ecommerce-dashboard
Tech: Power BI, DAX, Data Modeling
Real-time BI dashboard with live weather APIs, interactive filtering, and custom visualization design.
Built as an end-to-end API to BI pipeline project.
Tech: Power BI, APIs, Power Query, Figma, DAX
- Business Intelligence:
Power BI,DAX,Excel - Databases:
PostgreSQL,SQL - Automation and ETL:
Python,Power Query - Machine Learning:
Scikit-learn,XGBoost - Web and Data Products:
TypeScript,React - Tooling:
Git,GitHub,Bash
- Building production-grade dashboards for finance and operations.
- Turning chaotic transaction data into clean decision systems.
- Shipping full BI systems from raw data to executive dashboards.
- Combining ML, automation, and BI when it gives real business leverage.