balancing supply and demand

[OT from cottage renovations]

Spring is sprung and as the weather improves Capital Bikeshare usage is experiencing a large but predictable seasonal upswing, augmented by a recent fire sale promotion which doubled membership. 
There are now over 10,000 subscribers. 

The bikeshare system is completely addictive and over the winter was easy to use but it is now becoming increasingly difficult to find a bike or dock when needed and the system operator is having to scramble to keep up with demand. 

It all happens on the margin: each station without a bike is a trip that can’t start (or without an open dock is one that can’t end) representing a service failure for the next subscriber who comes along. 

Since system status is available as a near real time XML feed it didn't take much effort at all to put together a simple page showing bike and dock depletion updated hourly throughout the current day, as well as another one for the previous day’s performance.  If nothing else perhaps they can help subscribers and other stakeholders assess how the operator deals with the situation. 

like water, bikeshare bikes flow downhill

[OT from cottage renovations]

Rumour has it that Capital Bikeshare will soon be releasing systemwide usage data.  For those of us who derive amusement from such things the data will be interesting to play with analyze and may even yield useful insights. 

This could be quite timely as system usage is on the rise and demand is occasionally approaching or exceeding capacity, fueling much discussion among users and interested observers.  An assertion has been made that one cause of the distribution imbalance is a tendency of users to ride the bikes downhill and then choose some other mode for their return journey, leaving the bikes behind. 

A few days ago over on Housing Complex, Lydia DePillis released a preview containing station usage counts through 9 April.  Even in the absence of trip origin/destination and date/time detail, the usage totals can be used for a simple test of the "lazy user" assertion in the aggregate.  For each station: 

• compare trips in and out and determine net flow of movements in or out as percentage of total trips;
• determine elevation by feeding latitude/longitude to the SRTM elevation model via EarthTools web services.

---

The resulting chart is pretty straightforward, showing clearly that stations at higher elevations tend to be net exporters of bikes and the lower ones tend to be net importers. 

It goes to show that the bikeshare system is just like a railroad: it's necessary to haul those empties back up the hill for the next load.