Principal component analysis is a dimension reduction technique which allows visualizing the greatest variation in a data set. The first principal component (PC1) represents the greatest variation, followed by subsequent components (PC2, PC3,…).
This painting represents data points of a data set with fireworks. Since the first principle component displays the greatest variation in a data set, the fireworks are spread most horizontally. The second principle component (vertical, PC2), should not have as much variation as the first; therefore, the fireworks are not as spread out vertically as they are horizontally.