Long time R programmer here looking for greener pastures. What do clojurian data scientists use for a data.frame/data.table? My workflow so heavily depends on reading csv files into these structures. Usually on the order of >= 200k rows by 1k - 2k columns into memory where I can start to understand it interactively. Is this workflow smooth in clojure?

http://tech.ml might give you what you want, but seqs of maps and transducers work well too

H2o can be used from R or Java

in fact, the guy who wrote data.table, Matt Dowle, now works at H2o: https://www.linkedin.com/in/mattdowle