Business Analytics graduate student turning data into strategy — I work across SQL, Python, R, and AI frameworks to translate complex technical problems into business-ready solutions. I enjoy bridging the gap between engineering, product, and business teams, and I'm actively building toward AWS Cloud certification to help drive enterprise cloud adoption.

• 🎓 I'm currently pursuing my M.S. in Business Analytics at Lehigh University (Dec 2026)
• I'm currently learning toward AWS Certified Cloud Practitioner, AI Practitioner, and Solutions Architect Associate
• I'm looking to collaborate on Generative AI, NLP, and data analytics projects
• Ask me about SQL, Python, R, Tableau, Power BI, LLMs & Agentic AI
• 📫 How to reach me: tawandamatiashe2@gmail.com
• ⚡ Fun fact: Give me an aux cord and a road trip and I'll turn it into a whole event

Languages & Core Tech

Analytics & BI

AI & Data

Cloud & Tools

• Hospital Readmission Risk Audit : Predicts 30 day readmission on 99,340 hospital stays (AUC 0.671). The top 15% flagged group is readmitted 2.1 times as often as average, with a fairness audit by race and age and a hospital layer from CMS data.
• ClaimReady : Prior authorization and claims denial prevention agent for healthcare insurance, with a checked AI writer.
• FlowMind Process Mining (: Process mining on the Road Traffic Fine event log, with an AI investigator whose findings are checked against the data before they are shown.
• Agentic EHR Pre-Visit Orchestrator : Healthcare AI agent that prepares a chart summary before a visit, built on synthetic FHIR data.
• Chronic Disease Data Agent: Natural language agent using Google's Gemini API and Streamlit that turns plain English questions into structured queries on CDC health data, returning real answers and charts.

• Consumer Review Sentiment Analysis & Brand Benchmarking (NLP) — Evaluated 1,500+ unstructured reviews using R, VADER sentiment analysis, and NLP to benchmark L'Occitane and La Mer on customer satisfaction.
• Chronic Disease Data Agent — Built a natural-language AI agent using Google's Gemini API and Streamlit that turns plain-English questions into structured queries against CDC health data, returning real answers and charts.
• Predicting Online Book Club Subscriptions — Built linear and logistic regression models on 50,000 customers to predict subscription likelihood, with decile-based targeting for an email marketing campaign.
• Predicting Fraudulent Credit Card Charges with Machine Learning — Identified fraud in a 1.8M transaction dataset with a severe 0.52% class imbalance, using SMOTE and engineered temporal/geographic features to evaluate logistic regression and other models.
• Email Marketing Subscriber Churn & Lifecycle Analysis — Analyzed 30,000+ records with SQL and Python to map a 148-day churn window, projecting a 12–18% lift in retention.