a. Use LDA to classify the dataset into few classes so that at least 90% of information of dataset is explained through new classification. (Hint: model the variable qtr to variables togo, kicker, and ydline). How many LDs do you choose? Explain the reason.

Apply PCA, and identify the important principle components involving at least 90% of dataset variation. Explain your decision strategy? Plot principle components versus their variance.

Please provide the python code.