[QST] Streaming Integration and Catalog Based Dashboarding about cuxfilter HOT 1 CLOSED

Thanks @archenroot you have big plans!

If I understand your first question correctly: once a dataframe is loaded into GPU memory, it is accessible from anywhere. For example, if you load a cuDF dataframe from one notebook you can access it via a separate notebook. We don't have any direct visualization integration with Spark/Dask per say, but cuDF does have dask_cuDF as part of its offering and RAPIDS is actively developing closer Spark integration. There is also a cuDF cuStreamz library that might also be helpful.

For your second question: We don't plan on making an OmniSci level application with drag and drop type interface as our focus right now is more based on notebooks for the "interface" to creating dashboards. However, the updated cuXfilter does make it ALMOST drag and drop.

Feel free to keep asking questions, we are open to hearing feature ideas.

from cuxfilter.