This example shows data pre-processing procedure (dimension reduction) for a handwritten digit recognition project. Data used in this example is from the MNIST database of handwritten digits. 10000 images of digits and their corresponding values are provided. Principal Component Analysis (PCA) is first performed on the image data to reduce the number of features to 50. t_SNE method is then applied to further reduce the feature number to 2. The result is then plotted and shows apparent clustering.