Data Science Portfolio

Validated normality through production lot testing lots. Generated a test grid of simulated populations by mean and standard deviation of a key product dimension. Projected failure rate for simulated populations to guide improvements to part and process design.

Queried enterprise SQL database, filtered and grouped data on key part numbers. Illustrated deviation from expected process behavior, tracked performance over time, and validated adjustments to manufacturing process and equipment.

Built model to predict music genre from select features. Visualized predicted genre across two key features while holding remaining features steady at median values by creating a test grid to illustrate topology of predicted genre by prevalence of speech and instrumentals in a given song.

Used quality inspection data to visualize product deemed "OK to Ship" and drive investigations into outlying samples and improvements to process parameters.

Standardized and transformed data for modeling and classification.

Forecasted monthly data using Holt-Winters and ARIMA methods (including hyperparameter tuning).

Cross-validated model selection and tuning to produce best fit and honest assessment of model performance.

Performed exploratory data analysis and reviewed correlation matrix for model feature selection.