https://github.com/gitelman/BigData/blob/master/Project1/01-get-data.R
start small
- there's no point reading in a million rows if you can't read in 10 without error.
- there's no point reading in (50 states)/(10 years) if you haven't figured out what to do with one.
read function documentation: functions are often written to do sensible things if you don't specify certain arguments, if you are specific about arguments the function doesn't have to waste time figuring out what is sensible
look outside R - a tool specific to the job you need to do is often faster
split the problem up into "small" pieces