Analyzing Divvy

Chicago's bike sharing program, Divvy, recently published trip data for the 750,000 trips taken in 2013 as part of the "Divvy Data Challenge." They're looking to use data to answer burning questions like Where are riders going? When are they going there? How far do they ride? What are top stations? What interesting usage patterns emerge? What can the data reveal about how Chicago gets around on Divvy?

This page sets out to help answer some of those questions.



The Challenge My Approach

Insights

Metric What It Could Mean
Total Road Miles Biked: 1,260,127

Note: "Road Miles" do not take into account miles driven by riders who return bikes to the same station (round trips).
If you assume an average fuel efficiency of 24.9 miles per gallon, you could estimate that Divvy bikes saved approximately 50,600 gallons of gas in 2013. Granted, this grossly oversimplifies the correlation between bikes and cars as 1-to-1, and suggestions for a better calculation would be appreciated. According to the EPA's Greenhouse Gas Equivalencies Calcuator, there's enough carbon in 50,600 gallons of gasoline to provide a year's worth of energy for 22 homes.


Divvy rides averaged 1.6 miles per trip. Bikers are using Divvy to travel significant distances, not just around the block. A 1.6 mile trip spans about 13 blocks, or the equivalent of a leisurely ride from Millennium Park to Greektown. By cab, that's at least $7 (the price of a Divvy rental) before tip.

Round Trips account for only 5.54% of all Divvy rides. Divvy bikes aren't just for tourists. The overwhelming majority of Divvy trips took their riders from Point A to Point B, proving the model of bike sharing as another viable mode of public transportation.
Longest single Divvy ride: 17.6 miles One rider took a Divvy bike from the University of Chicago Medical Center in Hyde Park all the way to the north side intersection of Lincoln and Eastwood. If this route took Lakeshore Trail (as many online maps would suggest), they would have had an incredible view of Lake Michigan.

This trip actually happened twice in 2013: once around 10am on July 16th (a pair of riders), and again at 9pm on August 2nd. The high temperature on July 16th reached 93° with zero precipitation. On August 2nd, temperatures only reached 84° with 0.41" of precipitation.

Total Mileage of All Daily Trips, June to December 2013

Not terribly surprising: the total daily miles biked tend to peak in late summer (mid-September) and fall as the seasons begin to turn. The paired spikes in the data represent the weekends. Divvy bikers rode roughly 34% of total weekly miles on Saturdays and Sundays. This represents a 31.1% increase of miles ridden on the weekend compared to weekdays. For 2013, Divvy riders put in 826,100 miles during the week and 433,151 on the weekends. The single greatest day for total mileage was Saturday, September 28th, when Divvy riders trekked 18,577 miles.


(Click chart for interactive version)

Hourly Mileage, Week of Labor Day (Sept. 2 - Sept. 6, 2013)

The interesting piece is the consistent trilogy of peaks at 9AM (morning rush), 12pm (lunch), and 5pm (evening rush). The evening rush consistently charts about twice that of the morning rush. Ridership tends to fall off in the late evening fairly consistently as most folks apparently don't like to ride in the dark.


(Click chart for interactive version)

Total Mileage of All Daily Trips vs Daily High Temperature, June to December 2013

It would appear that many view 60 °F as their lower limit for outdoor cycling. The optimal outdoor temperature range runs from a breezy 60 °F to a toasty 90 °F. The plots at both of those endpoints appear roughly the same, and nearly doubles when the outdoor temperature falls between them. While it might appear linear, the relationship between daily total miles and outside air temperature doesn't represent a truly linear relationship. A regression analysis of the 71 days where the high temperature fell below 60 °F showed a correlation of roughly 0.815 and an R-square of 0.664.


(Click chart for interactive version)

Total Mileage of All Daily Trips vs Precipitation, June to December 2013

It can rain about 0.3" before a majority of folks avoid riding bikes. A few appear to actually enjoy a Divvy Downpour. Of the six months analyzed, only 14 days (two weeks) saw rainfall totals above 0.3", and some riders still took advantage of the Divvy system. On the rainiest day analyzed - October 31st (Halloween), AccuWeather reports 2.03" of rain fell on Chicago. Divvy riders persevered, however, and still logged approximately 2,000 miles that day.


(Click chart for interactive version)

Credits

Time Series Data hosted by TempoDB

Mileages calculated by MapQuest and © OpenStreetMap contributors

Weather data courtesy of AccuWeather
Reports from: CHICAGO/MIDWAY ARPT, IL [MDW] (Lat: 41.74 Lon:-87.78)

Charting courtesy of Pygal