Oh my god, am I a heretic among the believers? Let me explain why:
- Geekification might enhance the importance in the engineering culture of Silcon Valley, but where the technologists need to find jobs (i.e. enterprises), it only adds to the perception that this is too hard, why bother hiring such geeks?
- It will also distract the community from what I think is a fundamental commandment, "Thou shall make the technologies more consumable." If we continue to foster the concept that this set of capabilities are built by scientists, for scientists, we will lose sight of the fact that they should be built by scientists and engineers and interface folks and domain experts and ... for scientists and engineers and interface folks and domain experts and business people.
- It is an un-necessary rebranding of well know disciplines. I do not buy the fact that magically some new skills have to appear. Yes sure, hadoop is different. But skilled people grock new stuff using their existing repertoire of the basic skill of grocking large scale systems, analytics etc.
Am I too grumpy today? :)
In a word, yes. ;-)
No amount of pretending complex analytics are "easy" will make them so. The skills are not newly needed - they always have been. We've pursued "easy" for decades and gotten better and better at it for things everyday business users and analysts understand. But the hard stuff is still hard - because choosing the right statistical analysis, predictive model technique, or what have you requires depth. Tools don't fix that.
Posted by: Itmarketstrategy | July 06, 2011 at 04:32 PM
I agree with you. I also don't like the term data scientist. This will just make for more #hadoop black magic priest & less wide adoption by today data analysts.
Posted by: Racoss | July 11, 2011 at 03:46 PM
Yet another great post. As with past high end systems, eventually big data analysis will be done by regular business analysts and so I think your observations are right. And great to see you blogging regularly.
Posted by: Anshu Sharma | July 13, 2011 at 02:55 PM
The key ingredient for building predictive models is domain expertise, because that's the only thing that helps you to find meaningful predictive variables. I fully agree that we don't need more "scientists" as intermediaries between business and its problems. @itmarketstrategy: with the advent of ensemble modeling techniques, built on "best of breed" data models, tools can indeed help finding "the right" modeling approach...
Posted by: Alexander Lang | July 19, 2011 at 02:50 PM
A rose by any other name still smells as sweet.
I think most organizations, in the near term, will be scared off more by the pricetag of a data scientist than the name itself.
Working with Hadoop and other schemaless data formats isn't cheap, and is still outside the reach of a lot of organizations that could make use of the technology.
Still, we saw this as late as the late '90s, where it took four full-time DBAs getting paid north of $80-100k to manage a single e-commerce website (running TCL and an early SQL variant) for Nokia during my time there.
Posted by: Rizzn | November 03, 2011 at 11:58 AM
great post, very informative
http://sevenit.com/Infrastructure-Solutions.php
Posted by: Account Deleted | January 19, 2012 at 02:55 AM