The smallest number reported in z stats tables is (typically) -3, which gives a p value p <0.0002 and is smaller than my chosen significance level alpha=0.05, so it is statistically significant. We reject the null hypothesis and accept alternative hypothesis that the proportion of subscribers biking on weekends is less than the proportion of customers biking on weekends. I also check the robustness of my answer by using 201507 Citi Bike data, and the z statistic is even larger. This shows my hypothesis is robust in the summer.
GitHub Link: