* fast but not always * join ("SQL-like") * ml data science * datatypes (datetime as index) * very-big tables (dask dataframe)
