This shows the usage pattern of each user type on each day of the week. Qualitatively, it can be seen that the subscriber usage is relatively higher during the week, as opposed to on the weekend, whereas customer usage is relatively higher over the weekend.

Methodology

The overall counts for customers and subscribers were taken for each day of the week and were subsequently aggregated into counts for the week and the weekend. The null and alternative hypotheses are as follows:
\(H_0:\ \frac{P_{cus,wkd}}{P_{cus,wk}}<=\frac{P_{sub,wkd}}{P_{sub,wk}}\)
\(H_A:\ \frac{P_{cus,wkd}}{P_{cus,wk}}>\frac{P_{sub,wkd}}{P_{sub,wk}}\)
where \(P\) is the count for a user type during either the week or the weekend. \(cus\) and \(sub\) represent customer and subscriber respectively and \(wkd\) and \(wk\) signify weekend and week respectively. Pearson's Chi-Squared test was chosen to test the significance of the difference between the frequencies of two categories. A significance level of \(\alpha=0.05\) was used. Alternatively, a Z test could be used to test the proportions.

Conclusions

The counts by user type for each day of the week are given in Table 1.