I absolute loved the point made by Matthew Napleton - we must make tools easier to use!
That doesn't mean we eliminate the need for human interaction. As long as a decision has to be made, a machine can get you further, faster but the magic still happens when a person applies their intellect to the process.
Before leaping to the notion that automation is going to take our jobs (no Southpark reference intended), I think about all of the other areas of science that have benefited from automation: lab technicians don't have to shake, stir or rock vials of blood for ten minutes before doing their analysis. There is a tool for that portion of the program. Or meteorologists. Yes, you can still get in your van and chase tornadoes but now you have lots of other tools to help you analyze weather patterns.
What if we could HELP data scientists reserve their power by automating the steps prior to analytics? Let's start by automating the mundane tasks, the time-consuming, painful steps, the ones that create the most risk if missed, the ones that no one wants to do, or that no two people do the same (but should). Let's make data quality something easier to do. Let's make data integration a snap. Let's find ways to reuse our data work where possible.
That leaves more time for curiosity, for hypotheses, for the love of data science.
Cari Jaquet is VP of Marketing at Paxata.