Data science

>data science

I can't even put into words how fucking angry this meme makes me

for some cases looking at the data is really helpful like retail operations or looking at traffic flows etc. the techniques and processes are really refined and the people working with the data know how to make sense of it.
campaigns are far more complicated

Data scientist:
>statistician/mathematician with technology, programming, and business analytics experience

"Data scientist":
>23-year-old dude who just moved to San Francisco and knows how to install MongoDB and Cassandra

>muh big dahtah

>muh super innovative board room circle jerk

"Data science" is basically

1. Statistical knowledge
2. Econometric knowledge
3. Enough model-selection training to deal with big-k data (many features; "wide" data)
4. Enough programming knowledge to deal with big-N data (more observations than can fit in memory)
5. Enough training in scientific computing to be able to write numerically accurate software.

The good news is that (4) means you can safely apply asymptotic theory. Big-N is a storage/retrieval problem, not a statistical problem.

The bad news is that nobody is good at (3).

The worse news is that most of the data will be observational data, and almost no self-proclaimed "data scientists" have enough training in (2) to be trustworthy with observational data. No amount of statistics will help with this; you must have an econometric toolkit. Few people appreciate this point.

Could you stop pretending to be an expert on things you don't understand

I know that's asking a lot on Sup Forums

"Data Science is statistics on a Mac."

Most accurate description I ever read.

when will I be able to get a degree in Meme Science?

>tfw meme scientist and make a cool 80k a year


lol

you are a pioneer of meme science.

>tfw born 10 years too early to get a phd in memes
>tfw i will never work as a meme consultant for a major news outlet

;_;

who?

>tfw javascript codemonkey and make 160k a year

How much bullshit did you spew in the interview

believe it or not, some of us have jobs

yeah it's more of data math

Most topics in CS are math and not science.

>computer science
>not science
next thing you'll say it is not computer either...

Nope.

memes are more art than science

...

It's computer in the sense of computation, not the machines or the classic computers who were people that performed calculations for mathematicians and engineers. The "science" of computer science is in studying procedures, but because many of these procedures can be mathematically proven it's not a science. There aren't any possible theories of, for instance, bubble sort being able to run in O(logn), it's been proven mathematically to be O(n^2).

>Racist sexist startuper making 320k a year firing all except white males. It solves my problems and the binus problems.

yeah it's such a fucking meme ""profession""

You forgot the requirement of domain knowledge.

I've been watching "data scientists" embarrass themselves for years now because they have none

post taxes/check stub faggot

I work in Healthcare Analytics and the dependence statisticians have on ultra wide data kills me. Keep the relationships and abstract them away instead. Maybe their tooling doesn't allow it or maybe since they don't have a CS background they struggle with relationships. I don't know.

Sure, I get that it's easier to build and train models with a wide data set but when you're approaching week long run times and near a petabyte of flat data, there has to be a better way.

man, econometrics is so important, but it's easily the thing people forget about this field.

Though, I think you can be forgiven in not recognizing the differences between econometric and statistical practice, the literatures/perspectives are so different.

>mfw I'm an econ major who went into data science instead of a computer information systems guy who tried to read about econ