Joshua Scantlebury

Logo

View My GitHub Profile

Joshua Scantlebury

Data Analytics | Data Science
Passionate about transforming data into actionable insights

πŸ”§ Tech Stack: Python, SQL, R, Scikit-learn, TensorFlow/PyTorch, Gurobi, Tableau, PowerBI, Pandas, NumPy, AWS, Git, Bayesian Analysis, NLP, Time Series Forecasting, Optimization Models


πŸš€ Projects

Predicting NBA Player Performance

Python, NBA API, Scikit-learn, Ridge Regression, Feature Engineering
Developed a model using public NBA data to predict player scoring, outperforming simple averages by capturing trends via rest days, home-court advantage, and rolling stats.


Data-Driven Marketing for Shine Inc.

Python, RFM Analysis, Market Basket Analysis, ARIMA
Boosted sales forecasts by $32K/month for a cosmetics company by identifying high-value customer segments and optimizing product bundles.


Multiclass Classification Optimization

Python, Random Forest, Bayesian Analysis, PCA
Tackled class imbalance and feature correlation, achieving a 6.25% F1-score improvement using feature-selected Random Forests and Bayesian uncertainty estimation.


Clinical Trial Labeling Automation

Python, Transformers (Hugging Face), NLP, Multi-Label Classification
Built a deep learning model to auto-label clinical trial termination reasons, reducing manual curation time by 40% in data-rich cases.


Automating High-Impact Article Selection

Python, Random Forest, NLP, Scikit-learn
Automated article curation with 66% accuracy, targeting content likely to exceed 1400 shares for 25% higher engagement.


Fraud Detection at AJOS Bank

Python, Random Forest, SMOTE, Imbalanced Learning
Reduced fraud losses with a 99.9% accurate model by balancing data (60:40 fraud ratio) and optimizing feature selection.


Spare Parts Logistics Optimization

Python, Gurobi, Pandas, Operations Research
Cut costs by $150K+ via a depot allocation model with real-world constraints (supplier limits, capacity, service levels).


NYC Airbnb Pricing Strategy

Tableau, Market Analysis, Data Visualization
Delivered a dashboard guiding pricing decisions ($190/night benchmark) and renovation ROI analysis.
Dashboard


🌟 Interests


πŸ“¬ Let’s Connect!

Open to collaborations, data challenges, and geeking out over the latest in ML!