In order to download the zipped data from http://www.transtats.bts.gov/, run
python download_data.py
10 years of data will be roughly ~3GB with each month being around 22.1MB large zipped.
To download individual months pass a list of tuples (year, month)
to download_zip_files
dates = [(2014, 1), (2014, 2)]
download_zip_files(dates, urldict, cache_path)
As retrieving the file urls takes some time, they are preparsed into a dict where keys adressing year/month are encoded as "201402" for the file url of 2014/02.