Josh Gammon

Built Decision Tree and Random Forest classification models to classify skin disorders. Focused on the full ML pipeline: exploratory data analysis, model training, and evaluation.

• Performed preprocessing, encoding, standardization and model training on a dermatology dataset.
• Trained Decision Tree and Random Forest classifiers and evaluated with accuracy, precision, and recall.
• Used confusion matrix and feature importance plots to interpret model behavior.

Tech: Python, scikit-learn, pandas, matplotlib, Jupyter

Built models to predict whether it will rain using historical weather data. Compared Logistic Regression and Support Vector Machines to understand the tradeoffs between a simple linear model and a more flexible classifier.

• Cleaned and normalized features such as temperature, humidity, and rainfall.
• Trained Logistic Regression and SVM classifiers for a rain / no-rain target.
• Explored various kernel methods and used Grid Search to find optimal hyperparameters.

Tech: Python, scikit-learn, pandas, NumPy, matplotlib, Jupyter

Analyzed county-level COVID-19 data for Alabama to explore death rates, vaccination rates, and provider distribution. Engineered metrics such as mortality rate and vaccine coverage, then visualized trends to identify high-risk regions.

• Cleaned and prepared multi-source datasets using pandas.
• Calculated new features such as Death Rate (%) and Vaccination Rate per 100k.
• Built bar and heatmap visualizations for insights across 67 counties.

Tech: Python, pandas, matplotlib, seaborn, Jupyter

Used NLTK to analyze a corpus of Shakespeare’s plays, from tokenization and cleaning through word frequency and analysis. Compared vocabulary and phrase usage to get a feel for how style differs across works.

• Tokenized text, removed stopwords, and normalized tokens.
• Computed word and bigram frequencies to identify recurring patterns.
• Assorted words alphabetically and removed duplicates to show unique words.

Tech: Python, NLTK, pandas, matplotlib, Jupyter