Data science + decision modeling experience across machine learning (Python), optimization/simulation (Excel), and BI reporting (SQL/Tableau).
Predictive Modeling + Machine Learning (Python)
Data Science for Business: Technical (NYU Stern | Spring 2025)
Applied data science: built end-to-end predictive analytics skills in Python—from EDA and feature engineering to machine learning models and rigorous evaluation—and delivered a team capstone using real operational data.
What I built
- Data science workflow (problem → data → insight): framed business questions as prediction tasks, performed data cleaning + exploratory data analysis (EDA), and created baseline comparisons to keep modeling grounded.
- Predictive modeling (regression + classification): implemented and compared models including decision trees, linear regression, ridge/lasso regularization, and gradient boosting, using train/validation/test splits and cross-validation to reduce overfitting.
- Model evaluation + business tradeoffs: evaluated performance with RMSE/MAE/R² (regression) and confusion matrix, precision/recall, ROC-AUC, lift (classification), connecting threshold choice to cost/benefit decisions (e.g., churn targeting).
- Feature engineering + representation: engineered features and prepared data with one-hot encoding, scaling/normalization, and transformations; worked with text features (TF-IDF) and similarity/kNN/clustering concepts for segmentation-style problems.
Tools
Python (Jupyter/Notebooks), pandas, NumPy, scikit-learn, matplotlib; model workflows including train_test_split, StandardScaler, Grid Search / Cross-Validation, and metrics for classification + regression.
Keywords
Python • Exploratory Data Analysis (EDA) • Feature Engineering • Decision Trees • Regression • Ridge / Lasso • Gradient Boosting • Model Evaluation • Cross-Validation • RMSE / R² • ROC-AUC / Precision-Recall • TF-IDF • kNN / Clustering
Optimization + Simulation Modeling (Excel Solver)
Decision Models & Analytics (NYU Stern | Fall 2023)
Prescriptive analytics: built decision models that minimize cost / maximize revenue and quantify risk under uncertainty using Excel-based optimization and Monte Carlo simulation.
What I built
- Optimization models (LP / Integer / Nonlinear): translated business problems into decision variables, objectives, and constraints, then solved with Solver (Simplex LP + integer + GRG/Evolutionary when needed).
- Constraint design + model engineering: implemented scalable spreadsheet logic (e.g., SUMPRODUCT objectives, SUMIF aggregation, weighted-average constraints, matrix math) to keep models auditable and easy to modify.
- Sensitivity & what-if analysis: interpreted Solver outputs (e.g., shadow price / reduced cost / allowable ranges) to answer “what changes matter?” and communicate tradeoffs.
- Simulation & risk: built Monte Carlo models with appropriate probability distributions, summarized outcomes using percentiles/probabilities/confidence intervals, and (when applicable) combined simulation with optimization to select decisions under risk limits.
Tools
Excel Solver (Simplex LP / GRG / Evolutionary), OpenSolver, Crystal Ball, OptQuest, Excel modeling (SUMPRODUCT, SUMIF, MMULT, weighted averages), charting + graphs for distributions/histograms.
Keywords
Optimization • Linear Programming • Integer Programming • Nonlinear Optimization • Sensitivity Analysis • Monte Carlo Simulation • Decision Modeling
BI Reporting + Dashboards (SQL)
Information Technology in Business & Society (NYU Stern | Spring 2023)
Business-tech analytics: pulled insights with SQL and communicated them through Tableau dashboards—grounded in data quality, governance, and security/ risk awareness.
What I built
- SQL for business reporting: wrote queries to filter/segment records, aggregate metrics (COUNT/SUM/AVG), and join tables to produce reporting-style outputs that answer stakeholder questions.
- Tableau dashboards + BI storytelling: connected multi-table datasets, built a clean data model (relationships/joins), created calculated fields, and delivered dashboards with comparison views, trends, and table calculations (rank / percent-of-total / running totals).
- Data quality & governance mindset: identified common “dirty data” causes and applied practical cleanup approaches (rules, validation, automation/tagging) so insights remain trustworthy.
- Security + implementation awareness: understood SDLC vs Agile tradeoffs and core information security goals (confidentiality / integrity / availability), so analytics work is grounded in real operational constraints.
Tools
SQL (PostgreSQL/pgAdmin), Tableau Desktop (data modeling + dashboards), Excel/Google Sheets (analysis + reporting)
Keywords
Information Systems Strategy • SQL • Tableau • BI Dashboards • Data Modeling/Joins • Data Quality • Governance • SDLC/Agile • Security & Risk
Automation + Data Processing (Python)
Introduction to Programming and Data Science (NYU | Spring 2023)
Programming: built Python skills to write clean, reusable code for data tasks—covering control flow, functions, debugging, data structures, and file I/O.
What I built
- Core Python + control flow + debugging: practiced types/strings, common errors, if-else, and loops to solve structured problems; handled edge cases and used break/continue.
- Functions + program design: wrote reusable functions and used flowcharts to reason about execution order; kept I/O (input/print) outside core logic for easier testing.
- Data structures for data work: used lists / sets / dictionaries (including nested structures) to organize data and support loop-based processing (e.g., counting / “histogram” patterns).
- Persistence + File I/O: read/write files and converted structured text (e.g., TSV) into Python-friendly structures; used lightweight notebook workflows for file navigation and inspection.
Tools
Python (Jupyter / notebooks), file I/O, and structured problem-solving methods (flowcharts, debugging workflow, iteration patterns).
Keywords
Python • Functions • Control Flow (If/Loops) • Debugging • Data Structures (Lists/Sets/Dicts) • Nested Structures • File I/O • Flowcharts
Tong Zhao 