THE RESULT
What the data shows
The largest computed metric is 0.4107 for 2012-02; the smallest is 0.3652 for 2011-07. Metric: mean_peak_share (Share of daily rentals).
Busiest hours are selected retrospectively for each day. This is not an advance peak-hour forecast.
| month | complete days | mean peak share | median peak share |
|---|---|---|---|
| 2011-01 | 11 | 0.4084 | 0.4091 |
| 2011-02 | 12 | 0.3993 | 0.3992 |
| 2011-03 | 18 | 0.4068 | 0.4091 |
| 2011-04 | 29 | 0.4098 | 0.4046 |
| 2011-05 | 31 | 0.3867 | 0.3919 |
| 2011-06 | 30 | 0.3824 | 0.3941 |
| 2011-07 | 31 | 0.3652 | 0.3695 |
| 2011-08 | 29 | 0.3934 | 0.4069 |
| 2011-09 | 27 | 0.39 | 0.3989 |
| 2011-10 | 30 | 0.3926 | 0.4017 |
| 2011-11 | 29 | 0.4013 | 0.4057 |
| 2011-12 | 28 | 0.3943 | 0.3938 |
THE METHOD
From source to answer
Daily concentration restricted to dates with all 24 observed hours.
The Python source exposes this study’s transformations. The complete project download includes shared preparation and evaluation routines.
THE NEXT DECISION
What follows from the finding
Evaluate peak service requirements with actual fleet and station constraints.
Where the conclusion stops
Historical system rentals measure realized use, not unmet demand or station inventory. Missing hours are unknown. Weather associations do not establish causal effects.
Related studies may reuse observations or holdouts. These are historical analyses; associations and backtests do not demonstrate commercial impact. Further model tuning needs new, untouched evaluation data.