The Hubway Data Visualization (#hubwayviz) is a challenge to take 500,000 data points and create something, anything. I don't have the top technical or design prowess, but I have had some fun with large data sets. Win, lose or draw, I hope these provide some insight in to shared bike mobility, or at least are of interest to someone.
I decided to begin the challenge by "playing" with the data. Make some charts, see what happens, see what's interesting. My first question was when bikes were in use, and by whom. Obviously there were differences in weekday and weekend use, and registered use and casual use, so I plotted out the daily use, and then refined it to show the whole system's use each week, in ten minute intervals.
This looks interesting. Huge peaks for registered users on weekdays, none for casual users. Registered users account for more then 90% of the total use at these peaks, and casual users see somewhat higher usage at lunch, evenings and especially weekends (and the not-at-all curious peaks right around 2 a.m. on Friday and Saturday nights). This would serve as the base of my analysis going forwards. I want to know who goes where, when, and at what speed.
So, I broke everything that I found in to several pages:
- A series of charts looking at different types an times of usage
- A series of maps showing speed and balance
- Usage for North and South Stations, the two busiest Hubway locations, which have dramatically different patterns
- A station-by-station look at the weekday usage patterns for registered users
- A quick analysis of my own Hubway use
- Some ideas of where someone (perhaps someone who can code better than I) could take this next