Another interesting point is that the number of users seems to show the seasonal effect in both group. After applying the dt.dayofweek function as shown in FiG. 6, it is interesting that the number of riders in Saturday and Sunday tend to less than that in weekdays for both two groups. In order to test whether there exist periodic time series, the Chi-Squared is applied. Under the null hypothesis that there is no seasonal effect (i.e., F0(x) is a uniform distribution) and significant level as 0.05, we get the p-value roughly 0.0 for both two samples, which is less than significant level 0.05. Therefore, the null hypothesis can be rejected and there exist seasonal effect in this dataset.