Collect data for an automobile's travel time between two locations with traffic factored in. traffic-student scrapes Google Maps to find the travel times in traffic between a user-defined point a and point b at the time of the request. Then, it dumps those travel times into two CSV files: one for a to b and one for b to a.
node scrape.js address_a address_b [output_file_a_to_b] [output_file_b_to_a]
The output files default to a_to_b.csv and b_to_a.csv respectively.
In order to get traffic-student running, you will first need to install Node.js and NPM if you don't have them already. The easiest place for you to get them is probably here. The installer should install both Node.js and NPM.
Run npm install in traffic-student's root directory.
traffic-student is intended to be run in a cron job. For example, on your Unix-like always-on box (you're on your own, Windows users), type crontab -e and enter the following line:
*/15 * * * * cd <traffic-student's root directory>; node scrape.js address_a address_b
Note: it's possible you'll need to give the full path the node if your cron isn't able to find the command. Run which node to find node's full path.
Ex: */15 * * * * cd <traffic-student's root directory>; /usr/local/bin/node scrape.js address_a address_b
If something's still not going right, you could always log standard error to an output file like so: <my screwed up crontab entry> 2> ~/Desktop/error.txt
Running this cron job will run traffic-student on the 00, 15, 30, and 45 minute marks of every hour of every day of the week. It will get the number of minutes of travel time in traffic* both directions and append those integers to the csv files. There are some prepared output files in traffic-student's root directory with column headers already set up.
Once you set up the cron job**, let traffic-student run for a few weeks, or even better, a few months. Then, drop the csv files into excel, and make some pretty line charts like the ones below.
x-axis: time of day, y-axis: travel time in minutes
from my home to work
from my work to home
These are the results from a few months' worth of traffic-student collecting data. I found these results pretty interesting. You may notice a few trends:
- a morning, lunch, and afternoon rush during weekdays
- traffic generally getting worse as the workweek goes on (Friday is significantly worse than Monday)
- largest lunch rush on Friday... makes sense
- going from work to home for me sucks relatively more if I go at peak times
Overall, this was really useful and definitely informs my decisions about how I make my commute. Hopefully it can do the same for you, too.
node scrape.js "215 N Lumpkin St Athens, GA 30601" "660 Peachtree St NE, Atlanta, GA 30308"
node scrape.js "215 N Lumpkin St Athens, GA 30601" "660 Peachtree St NE, Atlanta, GA 30308" my_custom_file1.csv my_custom_file2.csv
* Sometimes Google does not return the travel time in traffic at all. In this case, traffic-student will write a travel time of 0 to the csv. This generally happened between 1am and 5am for me, so I ended up cutting off the first six hours of each day's data on my results.
** Ideally, you'll want to do the initial kick off of your cron job after 11:45pm and before 12:00am. That way, you'll get a full row of values for your first day.