With the almost endless onslaught of new board games, how can someone discover the best and most diverse set of games without spending a boatload of cash?
After applying the run-of-the-mill k-means algorithm with a PCA eigendecomposition on a boardgamegeek.com dataset, I found six distinct(ish) clusters identified via the silhouette score. The chart below shows all six groupings accompanied with the highest rated games within each cluster.
