UO Bike Counter:

Visualizing Real-Time Sensor Data

Above is a graph showing bike count data for the University of Oregon. The bike counter was installed at the U of O entrance at 13th and University. The counters were installed under the paving tiles across the U of O entrance. The counters use an inductive loop counter, which uses an electromagnetic detection system to induce electrical current in a nearby wire. Unfortunately due to operator error, only data from the southern portion of the entrance is being collected.

Below is a Tableau story about the U of O bike count data.

Analysis and Observations
Based on the data there seemed to be a correlation between campus population (demand) and bicycle counts. When demand was high, bike counts were high as well (relatively speaking). I also noticed that during weekends bike and demand were significantly lower. Despite this strong correlation, when the data is broken down hourly we see discrepancies. The demand line follows more of a bell structure on most days, whereas the count data tends to fluctuate. The most obvious explanation is that counts drop during class times and increase between classes. When looking at the environmental data it was clear that it did not play a role in bicycle trip counts. At times (April 25) we see a correlation between percipitation levels and bicycle counts but by and large, this correlation is absent.

Sensor Description
The sensor used for collecting the data in this study is an induction/inductive loop counter. The counters use an inductive loop counter, which uses an electromagnetic detection system to induce electrical current in a nearby wire. When a bike travels over the coiled wire, the electromagnetic disturbance is measured. The bike counter was installed under the paver stones at the U of O entrance at the intersection of 13th and University.

Scale
I chose a daily count to show the correlation between demand and count data. An hourly scale was used to show discrepancies between demand and count data.