$[40$ pts$]$ Implement a correlation program from scratch to look at the correlations between

the features of Admission$\_$Predict.csv dataset file. $($This Graduate Admission dataset, with

$9$ features and $500$ data points, is not provided on Canvas; you have to download it from

Kaggle by following the instructions in the module Jupyter notebook.$)$ Remember, you are

not allowed to used numpy functions such as mean $(),$ stdev$(),$ cov $(),$ etc.

You may use DataFrame. corr $()$ only to verify the correctness of your from$-$scratch matrix.

Display the correlation matrix where each row and column are the features. $($Hint: this

should be an $8$ by $8$ matrix.$)$

Should we use 'serial no$\text{'}?$ Why or why not?

Observe that the diagonal of this matrix should have all $1\text{'}$s; why is this?

Since the last column can be used as the target $($dependent$)$ variable, what do you

think about the correlations between all the variables?

Which variable should be the most important to try to predict 'Chance of Admit'?