Data Analyst · Data Scientist
From Power BI dashboards to machine learning models, I turn raw data into insights that drive real decisions.
I am a Data Analyst and Data Scientist with hands-on experience in SQL, business intelligence, data engineering, and machine learning. Currently completing my MS in Data Analytics and Visualization at Yeshiva University's Katz School of Science and Health, I specialize in turning complex data into clear decisions through predictive modeling, interactive dashboards, and data storytelling. I have worked across insurtech, energy, and academia, delivering insights that drive measurable outcomes.
SnapRefund
Philadelphia, PA · Remote
Katz School of Science and Health
New York, NY
VEETNIC Enterprises
Hwange, Zimbabwe
Nicole's Treats
Harare, Zimbabwe
Zimbabwe Power Company (ZPC)
Hwange, Zimbabwe
Yeshiva University, Katz School
Jan 2025 to Dec 2026
University of Zimbabwe
Graduated Aug 2024
Interactive Power BI dashboard analyzing profitability across products, underwriter performance, and sales channels using simulated insurance data built for strategic decision making.
End-to-end sales performance dashboard transforming raw transactional data into interactive executive-level visuals and KPI tracking for Adventure Works.
dbt pipeline that turns raw, messy Medicare-style claims data into a star schema with 17 automated data quality tests and a dashboard-ready output for regional payment analysis.
End-to-end analysis of real-world airline data identifying delay and cancellation patterns using Python and SQL, visualized in Tableau for actionable travel insights.
Transformed OLTP health records into a full OLAP data warehouse using PostgreSQL and ETL pipelines to surface actionable chronic disease insights for urban health management.
Eleven end-to-end projects, from data cleaning and regression to random forests, XGBoost, and neural networks. Capstone: predicting household financial vulnerability from BLS survey data (PR-AUC 0.95, spending R² 0.97).
Machine learning model predicting equipment failures from sensor data comparing six algorithms and a stacking ensemble (98.5% accuracy, 88% failure precision) to reduce downtime and maintenance costs.
Classified airline customer tweets as positive, neutral, or negative, first with NLTK, TF-IDF, and Logistic Regression (80% accuracy), then by fine-tuning DistilBERT and deploying it as a live Gradio demo on Hugging Face.
I am actively seeking data analytics, data science, and business intelligence roles. If you have a project or opportunity that could use my skills, I would love to connect.