It's considered good practice to get a good look at your data for initial insights before digging into building your machine learning algorithm of choice. But how do you do that in the big-data era, where each point might have tens, if not hundreds of features? We'll introduce the current state of the art methods: PCA and TSNE. We'll take baby steps to slowly build intuition and understanding about how TSNE works and why is it so useful. Examples will be shown in Tensorboard to familiarize the audience with it.