Hey, thank you for the question, it’s really interesting!
- For now no, there’s no performance benefit to using Luna over R or Python. Once we get our current performance issues sorted out, Luna should outperform these, though admittedly that requires a lot of effort on our part.
- We truly admire the number and quality of libraries available in both languages – that’s why we’re planning on piggybacking that and providing Luna with the possibility to interface with both languages. Python is higher on the priorities list, but once we get to supporting it, adding R interoperability will be a breeze.
- Once we satisfy both of the above, things really start to get interesting. First of all, the interactive experience from Luna is much better than the best Python and R can provide (i.e. Jupyter notebooks). Having the data displayed in a visual flow encourages experimentation and tightens the feedback loop, effectively putting the user much closer to the data and allowing for a more natural workflow. Moreover, Luna’s visualization engine is much more extensible than that of Jupyter’s – we’re hoping to build a vast community around data visualizations alone, and given that they all use a standardized API, there will be much more choice and freedom when it comes to visualizing data.
We actually hope for Luna to be the logical conclusion for the current Jupyter + R/Python workflow, liberating users from the linear experience of writing out code sequentially line-by-line. Sequentially is not how humans think and we believe it is finally possible (and definitely desirable) to reflect that in the way we interact with computers and data.