This is a simple group of scripts to ultimately generate a dataset that has the driving distance (in km, according to Google Maps API) between every county population centroid in the lower 48 and each major U.S. city with over 2 million people in its primary statistical area.
Update (2014-09-13):
-
The dataset has been created and is available here.
-
For more information on the dataset and its contents, please see this README.
The R script create-db.r generates a database with one row per large city/county center combination in the lower 48 states.
The Python script (generate-driving-dist.py), per run, finds a row in the database that doesn't yet have a distance calculated, calculates the distance, and then saves it back to the database. It is meant to be run repetitively (as a cron job, for example).
Here is a sample cron tab:
* * * * * /path/to/python3 /path/to/generate-driving-dist.py /path/to/database.db >> /path/to/log
*/2 * * * * /path/to/python3 /path/to/generate-driving-dist.py /path/to/database/db >> /path/to/log
The final product will be a complete SQL database with all of the distance information. I will post the final database to this repo once it is complete.
Finally, the R script export-db.r creates the shortest_flag and then exports to the data to a CSV.
There are 102,597 combinations that need distances calculated. Google Maps API limits you to 2,500 queries a day, but because of the limitations of cron jobs, we will run 2160 per day. Therefore the estimated run time is 48 days.
Update (2014-09-13): The process began on 2014-07-04 and completed on 2014-08-21.
County Density/Distance/Population Map. Erickson, Jeffrey P. 2014. (link)