New Poll: Did you do Data Science for Humans or for Machines?

Did you do mainly Data Science for Humans (explainable models for a human decision maker), Data Science for Machines (accuracy-oriented black-box models for automated algorithms), or both? Please vote.

By Gregory Piatetsky, @kdnuggets, Sep 1, 2014.

This poll is now closed - here are the results of the poll:
Data Science is mainly a Human Science.

A recent article in TechRepublic suggested that there are 2 kinds of data scientists
Human vs Robot

- those that work on "Data science for humans", where the ultimate decision-maker and consumer of data analysis is a person, and where the models need to be understandable (even if less accurate) and packaged as part of a story.

- those that work on "Data science for machines", where the ultimate user is a computer (e.g. for ad targeting, product recommendations) and the goal is to build the most accurate model possible, with understandability less important.
This division is also roughly equivalent to data science consulting (human-oriented) vs products (autonomous, machine-oriented).

The article also claimed that data scientists usually are good in only one or the other type of analysis, and that managers should hire the right kind of Data Scientist for their task.

I don't quite believe that this division is so clear-cut. Although I have mostly worked on "human-oriented" data science, I have also done machine-oriented analysis, and think that good data scientists can do both.