Blog
Thoughts on data science, machine learning, and beyond.
K-Nearest Neighbours: how many neighbours should you actually trust?
A plain-English introduction to K-Nearest Neighbours, and why picking K is really a balancing act between overfitting and underfitting.
A brief history of data science
From least squares in the 19th century to Python and modern machine learning — a short tour of how data science became what it is today.
Predicting Titanic survival: my first Kaggle ML project
How I built a Gradient Boosting classifier to predict Titanic passenger survival, achieving a Kaggle score of 0.78, and what the feature analysis revealed about who was most likely to survive.
Predicting DVD rental duration with regression models
How I used Lasso feature selection and five regression models to predict how long a customer will rent a DVD, hitting an MSE of 2.50 against a target of 3.
Can unsupervised learning identify penguin species on its own?
Using K-Means and t-SNE on the Palmer Penguins dataset, I explored whether physical features alone can reveal species groupings — and how accounting for sexual dimorphism transformed the results.