Download glass data from the UCI machine learning data repository (https://archive.ics.uci.edu/ml/machine-learning-databases/glass/). The dataset is about glass identification based on its properties. The column headings in sequence from first to last are observation number, refractive index, sodium (content in weight percent), magnesium, aluminium, silicon, potassium, calcium, barium, iron and type of glass (last column). From the data provided, answer the questions below.

(a) Test for normality of all the variables separately using normal probability plot, except the first and last column variables. 2 marks

(b) Construct a 90% confidence interval for each of the variables in part (a) which can be considered as normally distributed separating the data between float glass and non-float glass. In the last column, the numbers 1 and 3 represent float glass, all other numbers indicate non-float glass. In other words, there will be two confidence interval constructions for each normally distributed variable - one for the float glass observations and the other for the non-float glass observations. After all the confidence intervals have been constructed, identify the variable(s) which are significantly different in composition between float and non-float type of glasses (i.e., identify the variables where the confidence intervals do not overlap).