I don't think the API is well suited to data analysis.
If you want to do analysis of the SE dataset(s), you should download the latest torrent file. Depending on how which, and how much, data you need for your analysis, you might be able to combine the dump with some judicious use of the API to pull current metadata from the site.