One $($cool$?)$ application of PCA is lossy$-$compression. We can project our data down to k dimenions

$($where k $<$ D$),$ then when we need to reconstruct our data, we can just multiply the k dimensional

data by $($the transpose of$)$ its k associated principle components in order to return to our original

feature space!

In this part of the assignment you$$ll be making a video showing how the reconstruction looks as you

varying k$,$ from k $=1$ to k $=$ D$.$ affects the reconstruction, visualizing this as a video. To create your

video, you can use Python$$s opencv module: pip$3$ install opencv$-$python. See the Resources section

in BBlearn for an example of Python commands that can create a video. In addition, you may want

to use numpy functions like reshape, uint$8.$