Following the links on the
program at the Georgia Tech
Web site, the program looks like
a fairly wide buffet from
practical computing, current
business applications,
statistics, and operations research.
Georgia Tech is especially strong
in operations research.
So, here data science is a new bottle of wine
blended from some now
quite well known old bottles
of wine. And it is not nearly the
first such blending since
there have also been programs
such as mathematical sciences and
applied mathematics. Other
blendings have included
mathematical finance, financial
engineering, and bio-statistics.
Apparently the high current interest
is because
now the associated computing
is much cheaper, more powerful,
and easier to use. And there has
been a lot of hype from some
sources.
However, I question if US
mainline
business is
much interested: IMHO and my
experience says that nearly any
specialized technical material
faces a serious obstacle since
in the organization chart the highest
ranking technical person (if not the
CEO then necessarily a subordinate)
has to report to a supervisor who knows from much less to
nearly nothing about what that
technical
subordinate person is doing.
MD doctors, CPA accountants,
licensed engineers, and
licensed lawyers have some
crucial, serious professional status,
processes, support, etc.
that is missing with applied
mathematicians, statisticians,
data scientists, etc.
For software developers, roughly,
the solution is for the organization
to have a CIO, all the developers
are in the CIO's organization so
report only to experienced
developers, and only the
CIO reports to, interfaces with,
non-experts in computing.
Computing is now so darned important
that the rest of the C-suite
has to swallow their pride and
accept the CIO at the table.
Net, I fear that data scientists
will have too little professional
or organizational protection
from rain falling down the
organization chart from the C-suite.
Or, for the supervisor, most
projects will be lose-lose:
If the project fails, then the
supervisor has a black mark
from wasting money on a failed
project. So, with a
failed project, the supervisor loses.
If the project is successful,
then the supervisor
and, maybe, everyone
in the C-suite, maybe even including
the CEO, can be afraid of the
project leader now regarded
as a 900 pound tiger and, thus,
a loss for the C-suite.
Here the organization chart
from the project leader up to the
CEO is engaging in classic
goal subordination, that is,
pursuing what is best for themselves
personally while
sacrificing what is good for the
company.
And for startups, what fraction of
venture partners would be able to
evaluate a proposal that makes
heavy use of some of the more
advanced applied math in that
Georgia Tech program? Net, the
venture partners don't know the
technical material, either.
Or, as I suggested, nearly all wine
in the blend is now quite old,
and it didn't achieve much
traction in mainline business.
My short summary view is that for
such technical material, especially
material more advanced than in the
Georgia Tech program, and for a
startup, the founder CEO needs to
be both (A) the main expert in the
technical material
and (B) essentially a solo founder
who can write the software, bring it
to market, and get the coveted
traction significantly high and
growing rapidly -- at which time
the founder may not be willing to accept
equity funding and report to a
BoD that does not understand the work,
that is, be back in the situation of
a technical subordinate reporting
to a supervisor who does not understand
the technical work and, with the low
expenses of a one person company,
just grow organically from
revenue.
Or, IMHO, the most promising career
future of an applied mathematician, etc., in business is to be a solo founder
of a startup.
Georgia Tech is especially strong in operations research.
So, here data science is a new bottle of wine blended from some now quite well known old bottles of wine. And it is not nearly the first such blending since there have also been programs such as mathematical sciences and applied mathematics. Other blendings have included mathematical finance, financial engineering, and bio-statistics.
Apparently the high current interest is because now the associated computing is much cheaper, more powerful, and easier to use. And there has been a lot of hype from some sources.
However, I question if US mainline business is much interested: IMHO and my experience says that nearly any specialized technical material faces a serious obstacle since in the organization chart the highest ranking technical person (if not the CEO then necessarily a subordinate) has to report to a supervisor who knows from much less to nearly nothing about what that technical subordinate person is doing.
MD doctors, CPA accountants, licensed engineers, and licensed lawyers have some crucial, serious professional status, processes, support, etc. that is missing with applied mathematicians, statisticians, data scientists, etc.
For software developers, roughly, the solution is for the organization to have a CIO, all the developers are in the CIO's organization so report only to experienced developers, and only the CIO reports to, interfaces with, non-experts in computing.
Computing is now so darned important that the rest of the C-suite has to swallow their pride and accept the CIO at the table.
Net, I fear that data scientists will have too little professional or organizational protection from rain falling down the organization chart from the C-suite.
Or, for the supervisor, most projects will be lose-lose: If the project fails, then the supervisor has a black mark from wasting money on a failed project. So, with a failed project, the supervisor loses.
If the project is successful, then the supervisor and, maybe, everyone in the C-suite, maybe even including the CEO, can be afraid of the project leader now regarded as a 900 pound tiger and, thus, a loss for the C-suite.
Here the organization chart from the project leader up to the CEO is engaging in classic goal subordination, that is, pursuing what is best for themselves personally while sacrificing what is good for the company.
And for startups, what fraction of venture partners would be able to evaluate a proposal that makes heavy use of some of the more advanced applied math in that Georgia Tech program? Net, the venture partners don't know the technical material, either.
Or, as I suggested, nearly all wine in the blend is now quite old, and it didn't achieve much traction in mainline business.
My short summary view is that for such technical material, especially material more advanced than in the Georgia Tech program, and for a startup, the founder CEO needs to be both (A) the main expert in the technical material and (B) essentially a solo founder who can write the software, bring it to market, and get the coveted traction significantly high and growing rapidly -- at which time the founder may not be willing to accept equity funding and report to a BoD that does not understand the work, that is, be back in the situation of a technical subordinate reporting to a supervisor who does not understand the technical work and, with the low expenses of a one person company, just grow organically from revenue.
Or, IMHO, the most promising career future of an applied mathematician, etc., in business is to be a solo founder of a startup.