I like that it's easy to use, has lots of options for modeling problems (especially when you have many features), is flexible enough so as not too restrictive in terms or what data sources can be used with your model - which makes things easier than if there was one fixed way we had always done everything! The only thing i dislike about rocketml sparse logit regression toolkit/library would definitely need improvement are some user interface issues such us making sure all input parameters were entered correctly before submitting them into an analysis request etc., but this stuff gets better over time since its still relatively new product compared to other existing software out their.