Back in part 1, we preprocessed our housing data and performed some EDA on it. We then proceeded to model the fixed, non-renovatable features through a series of regression analyses in part 2. [Read More]

With our dataset preprocessed from 81 features to the 44 that we have currently in part 1, we will now model our dataset with a variety of regression models to identify the best model to predict housing prices. [Read More]

Everyone knows about the Boston Housing Dataset. But I bet you might not have heard of the Ames, Iowa Housing dataset. It was featured as part of a Kaggle competition 2 years back and was significant in how it tested advanced regression techniques in the form of creative feature engineering... [Read More]

In part 1, we scraped for data off Boardgamegeek (BGG). Part 2 saw us making predictions off the ratings matrix with 5 models and evaluating offline through RMSE and top 20 lists recommended for me. [Read More]

In part 1, we extracted the games and ratings data via webscraping and api calls to the Boardgamegeek website. [Read More]