This is a project that started at an event put on by the Houston Data Visualization Meetup Group. The dataset was provided by Houston BCycle, a bike sharing system. I worked on the data using python and KeplerGL, an open source geospatial visualization tool built by Uber.
- Weather - cold and wet way worse than heat?
- How have usage patterns changed over time as stations were added
- Most popular station pairs
- Does use vary by local user zip codes vs visitors from far away zip codes
- are any patterns predictable
- no bikes or one bikes = lost revenue and frustrated customers could be mechanical
- full stations = 0 or 1 dock
- how much does a station being down cost
- usage patterns change by day of week
- fastest someone has ridden
- mininetworks are they islands eg rice, med-center
- super riders vs infrequent riders someone with an annual membership
- repeat single riders that might be better off with a membership 13/mo 79/yr
- How does weather - heat, cold, rain, etc. - affect ridership? Are some rider types, stations, areas, more or less affected?
- How have trip patterns changed as the system has expanded?
- What are the most popular trip pairs and what do they have in common? Distance, mini network, transit connection, something else?
- Where do members and users live? How does use of the system compare between locals and visitors: locations, trip types, times distances, etc.?
- How predictable are patters at individual stations. For example, do some stations always gain or lose bikes during certain time periods?
- How do usage patterns change by time of day or day of week or season?
- What's the fastest someone has ridden between stations?
- When and where will our 1,000,000th trip occur?
- How self-contained are the mini networks (Rice, UH, TSU, TMC) or how much do people connect to the rest of the network?
- How do the trip patterns of super users compare to others?
- Are there people who buy single rides repeatedly and what are their trip patterns?
- Follow an E-bike around
- E bike Numbers: 15040, 15050, 15052, 15062, 15074, 15056
8/14 and earlier has these unique fields:
Checkout Dock
Bike Checkout Method
Return Dock
Adjusted Charge
9/14 and after has these unique fields:
Trip ID
UserProgramName
UserRole
UserCity
UserState
UserCountry
Adjustment Flag
TripOver30Min
LocalProgramFlag
TripRouteCategory
TripProgramName
And both have the following (using / to indicate fields that are labeled differently, but mean the same thing):
User ID
Zip/UserZip
MembershipType
Bike
CheckoutDate/CheckoutDateLocal
CheckoutTime/CheckoutTimeLocal
CheckoutKiosk/CheckoutKioskName
ReturnDate/ReturnDateLocal
ReturnTime/ReturnTimeLocal-
ReturnKiosk/ReturnKioskName
Distance
EstimatedCarbonOffset
EstimatedCaloriesBurned
Duration
BikeModel/BikeType
Charge/UsageFee
AdjustedDuration