Checking Normality of Numerical and Categorical Data

Question

I have come across 3 questions on the title subject.

1. Why is it necessary to do a normality test? To check if data is imbalanced or not?

2. Are these 4 methods of checking if the data follows normal distribution criteria both applicable to numerical and categorical variable? I am trying to check if the data follows normal distribution by the following 4 methods.

   1. Checking Distribution
   2. Drawing Box Plot
   3. Drawing QQ Plot
   4. Use skewness, kurtosis criteria

3. Skewness for normal distribution is 0, Kurtosis for normal distribution is 3. Is there a certain bound that I can use to guarantee that the data is normally distributed? (such as, 0 +/- 1 OR 3 +/- 1)