Database Reference
In-Depth Information
How do you start a conversation with them about these techniques? You do
it by working in a tool that's familiar and comfortable with them, and then
you can slowly usher them into more modern tools. In the last chapter of my
book, I actually do introduce R and say, “Remember that forecast model that
we built that took us 50 pages in the topic and 10 tabs in Excel? It turns out
that you can call it in R in three lines.” I take people through that and show
them that you can just call the forecast function, and it just does it, which is
amazing. Could I have done that on page 1? Absolutely not. You can't do that
at the start. You've got to teach people first.
So that's the topic. I think it just goes back to my general philosophy, which
is that if you're going to do data science, you can't just buy the tools and sit
there. You have to know what tools are available to you. You have to know
the techniques intimately because you've worked through a book like mine.
Then when a problem comes along—wow, you'll actually have a toolbox that
doesn't just have a hammer in it. You'll have a toolbox that's got all of the
tools inside of it.
A lot of data scientists only do AI, whereas there are a lot of problems that
are solved and could be solved with optimization. There's even a whole range
of problems that can be solved with simulation. You've got to have all of these
tools in your toolbox. That way, when you know all these things, you can then
select the appropriate tool to solve the problem in a way that matters.
