Jeremy mentions a convolution matrix around minute 24 in his Introductory Lesson from the Data Institute. I wondered, how do you pick a matrix to use? Are there standard "effects"? This URL has some examples and a terrific online interactive demo of various "kernels".
I tried to post this to the Lesson 2 Forum, instead of in this general area, but either I'm too blind to see the New Post button, or I haven't accrued enough points to make such a post.
If the course creators find that this online demo is helpful enough, I encourage you to place it on the Lesson 2 Wiki.