Based on the announcement it looks like the standard data-science libraries (numpy, pandas, matplotlib, seaborne) will be available. The covers a lot of use cases. However I would really appreciate the ability to set the python interpreter to whatever I want. Then I could access the niche libraries I need (shout-out to JSBSim for open source flight dynamic modelling). Or even better, hooking up my organizations custom libraries we built for our own specific data analysis applications. Huge value add. Shame this likely won't be possible. Unless I've misunderstood some part of this.