Methodology

As seen above, I calculated the fraction of subscribers and customers on weekdays and weekends, and to test my hypothesis (in the Introduction) I used the chi-square test. I chose the chi-square test for proportions over the z test since it is non-parametric, which is more appropriate for my data set (which doesn't follow a gaussian distribution, as an assumption for the z test). This test was also recommended to me by Dr. Bianco who commented on my homework. In order to conduct the z test, I created the contingency table as seen in my python notebook (and in Fig 4).