Exhibit 1 Assumptions
This data has been smoothed to limit noise created by the tendency of riders to travel at “common” times, i.e., leaving from home or office on the hour or half-hour and arriving at entry stations 6-12 minutes later.
1.tssmooth ma smooth_rc = rider_count, window(2 1 2)
tssmooth constructs a new series in which each observation is an average of nearby observations in the original series. In this case, a uniformly weighted moving average is taken using the original observation and the two points to either side of that observation.
Before smoothing, small spikes appeared at the second 6-minute interval of every half-hour; i.e. 2:06 pm, 2:36 pm, …, 6:06 pm, 6:36 pm, etc.
The single exception for all Exhibit 1 stations occurred at 7 pm; the spike occurs on the hour rather than the following 6-12 minute period. Because this bump is distinguishable from the noise, I left that region unsmoothed.