|
|
|
|
|
|
by butteredpopcorn
2499 days ago
|
|
|
It’s possible that if you could spin up an open source relational database like Postgres or MySQL and load your data set, you could do everything but the models and viz themselves in SQL. SQL will handle 3 mil rows in a snap. If you are transforming the contents of the cells then R is potentially nicer, but if you are mainly summarizing, grouping, subsetting, etc SQL is fast and readable and you can “dump” a CSV that’s fast to read in R or Stata. Your R or Stata code will be really short, really just a load step, a viz / pretty tables step, and then a model run step. It’ll be easy to swap out languages according to the team/prof/project requirements. |
|
|