Projects
A collection of data analysis projects showcasing real business impact through strategic insights and actionable recommendations.
Problem: Telemarketing campaigns achieved an 11.27% conversion rate with no clear understanding of what drives performance variation across customers, channels, and engagement patterns.
Approach: Analyzed 41K+ campaign records in Python using segmentation, feature engineering, and comparative conversion analysis across customer attributes, contact strategy, and interaction intensity.
Insight: Conversion is primarily driven by engagement depth rather than customer profile. Long calls (48.61% vs 0.81%), cellular outreach (14.74% vs 5.21%), and prior success (65.11% vs 8.83%) dominate conversion likelihood, indicating strong efficiency gains from better targeting and higher-quality interactions.
Three product categories drive ~70% of total revenue
Goal: Optimize a USD 137.1 Million global portfolio by prioritizing high-performing national markets and managing a 69% regional concentration risk.
Process: Conducted a diagnostic analysis on 100 country-level aggregates to identify revenue drivers versus volume traps.
Key Insight: Sub-Saharan Africa and Europe serve as the primary revenue anchors (>50%), while the Online distribution model achieves a peak profit margin of 33.04%.
Fall and Trauma occurred up to 16-times more than other HAC measures
Goal: Identify which preventable conditions posed the greatest patient safety risk in 2025.
Process: Cleaned and analyzed CMS HAC data in MySQL, then exported results to Excel for stakeholder-ready visuals.
Key Insight: Falls and Trauma accounted for the overwhelming majority of HAC variation in 2025, with rates up to 16 per 1,000 discharges, far exceeding other measures.
A one-day look at new Zillow listings in 10 of the largest U.S. cities, visualized through Power BI.
What I Did: Extracted & analyzed 380 Zillow listings with Python and Power BI to reveal pricing and property trends.
Findings:Median price $439.5K, 74% single-family homes; condos led in value ($1,660/sqft). LA & NY skewed averages; Chicago & San Antonio affordable.
Outcome: An interactive dashboard providing actionable market insights.
Uncovered $2M revenue leakage from customer churn
Problem: StaffWise’s revenue growth stalled despite strong signups—nearly $2M at risk from churn & failed payments.
Findings: 84% churn for monthly-plan customers; long-term clients drove 70%+ of revenue loss.
Fix: Leveraged Power Query, DAX, and interactive dashboards to pinpoint leaks and prioritize retention strategies.
Cut my monthly supermarket bill by 5.16% without buying less
Problem: My monthly supermarket purchases hit KSh. 14,408, likely overpaying by shopping out of habit at Naivas.
Findings: Carrefour alone cut costs by 3.3%, while a mixed-store basket saved KSh 743 (5.16%), with 5 items driving 77.65% of savings.
Fix: Used Excel, Power BI (DAX + dashboards) to compare store-level prices and recommend an optimal shopping strategy.