Effect of Snow on Trafficby Kristopher Erickson
Measuring the effect of snow on Atlanta traffic during two snow storms.
On January 28th a snowstorm hit Atlanta and paralyzed the city. It was followed by a second event which happened during the week of February 10. The city, and it's populace was unprepared for the first snowstorm causing multi-hour commutes for many. However Atlanta managed to avoid a second traffic catastrophe by closing schools and offices. I have been analyzing my personal Open Paths data and I would like to combine that with others and examine the traffic patterns of Atlantans during two snow events. A sample using my personal OpenPaths data can be found here: http://ericksondata.com/wp/2014/personal-traffic-patterns/
I intend to calculate average human velocity for each latitude/longitude bin (as can be seen in my previous work) then examine how those two snow days compared with average velocity during other weekdays. In addition I plan to use raw counts by bin to determine how many people were stranded on the interstates during the first snow day, compare that to how many people were off the interstate during the 2nd snow event, and then compared to normal weekdays.
Data will be used to calculate average velocity, bearing, location and raw count. This data will be aggregated to a specific latitude-longitude bin.
I have an encrypted hard drive, I can connect via network cable, and will work on this project offline. Secondly I will bin everything so that once metrics are calculated (avg bearing, avg velocity, avg distance, etc.) user identification will be completely eliminated and not accessible in any of the final output. I will limit data to the Atlanta area only. The corridor of interest is a square between these coordinates:
Furthermore the dates will be limited to January 1st through March 1st. Weekends can also be excluded. This will provide a good baseline to compare the snow days vs regular traffic days.
I plan on providing an interactive Tableau Public dashboard available to the public to utilize. Any data available will be limited to aggregate summary by time down to a latitude longitude bin. An example of the graphics (using my personal data) can be seen here: http://ericksondata.com/wp/2014/personal-traffic-patterns/