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:
https://github.com/louisalulu/PUI2018_ml4922/blob/master/HW8_ml4922/HW8_Assignment2_ml4922.ip