Hello again from the Cycle Atlanta Group!
(The Dream Team, minus Becca)
It is week five, and we are in the process of addressing some technical challenges that have become apparent in the project. In particular, we are using R to “trim the ends” of bike trip data. The data from the Cycle Atlanta app looks like strings of coordinates, representing the motion of the user at points in time during a trip. So, in the interest of both anonymization — we wouldn’t want to map a user’s trip to his house — and a more accurate picture of a “trip,” we are working out an algorithm to trim the ends of trips without losing too much valuable information. We have found that bicycle speeds, within individual saved trips, begin and end with long sequences of zeroes or values near zero. We think that this is because users do not start and stop recording trips while in motion (which makes intuitive sense).
We are also using R to remove signs of inaccurate signals within trip data. These “noisy” data points often result from weak GPS signals from users’ smartphones. Erratic signals make it difficult to accurately plot the path of a bike trip.
In addition, we are currently using several JavaScript visualization libraries: D3.js for interactive graphs and charts, Leaflet.js for map prototyping, and prospectively the DC.js visualization library to link graphs together for a final analytics dashboard.
For much of the backend work with the Cycle Atlanta database (testing and writing scripts, and adding tables) we are using MySQL Dashboard.
Dr. Tony Giarrusso of the Georgia Tech School of City and Regional Planning (and the College of Architecture) has been involved with DSSG as well, providing tutorials in ArcGIS and personal help on more than one project, including this one. We will need to use this tool to map trips to existing roads, instead of plotting the raw coordinates of trips.
And, of course, keep checking our WordPress site, beccabikes.wordpress.com, to get to know the local urban cycling community with us!