Colleges Graduate 10,000 This Year With Masters In Data Science Degrees (techtarget.com) 64
dcblogs writes: The Master of Science in Analytics was created in North Carolina State University in 2006. Today, there are about 280 colleges and universities that offer a similar graduate degree and in total, they will produce about 10,000 analytics master graduates in 2019. "The demand is there, but the supply [of data scientists] is catching up quickly," said Michael Rappa, who founded the Institute for Advanced Analytics at North Carolina State University. Graduates of these programs are typically called data scientists, a relatively new term that's often cited as one of the most in-demand occupations in the U.S. These programs aren't completely unique. Graduates with degrees in statistics, for instance, were forerunners of the shift to analytics. Despite the increase in graduates, the entry level salaries remain strong, typically beginning at $80K plus. Amazon recently cited data scientists as a second fastest internal growing occupations.
Analytics? (Score:1)
Isn't that what a computer is for?
I thought we already automated this shit
Eh, the "new" MBA and Poli Sci I guess
Re:Analytics? (Score:5, Insightful)
Probably. "Data Scientist" is not something that is very respected in certain circles. But some people apparently think that this is a good idea instead of getting a real statistician. A real statistician will be expensive, but at least will have an understanding of what numbers to find and what they mean. Probably the same people behind this madness that think coders need to be cheap because the ones you can get cheap are enough.
Re:Analytics? (Score:5, Informative)
It's not comparable to statistician. Here is what SMU focuses on for their MS in data science. Is some statistics, but with a whole lot of computer science.
https://datascience.smu.edu/ab... [smu.edu]
Statistical Foundations for Data Science - Experimental Design, Statistical Sampling, T-tests, Analysis of Variance, Linear Regression, Diagnostics and Checks for Statistical Methods, Interpretation and Communication of Results (both oral and written), Ethics of Statistical Analysis
Doing Data Science - Tools for Data Science, Reproducible Research, Data Selection, Data Wrangling, Exploratory Data Analysis (EDA), Machine Learning, Time Series Modeling and Forecasting
Applied Statistics: Inference and Modeling - Multiple Linear Regression and Variable Selection, Multivariate Analysis of Variance (MANOVA), Linear and Quadratic Discriminant Analysis, Unsupervised Learning (Clustering), Methods for Categorical Variables (Explanatory and Response), Autoregressive Models for Time Series Data, Basic Bootstrap
File Organization and Database Management - Database Queries, Relational Database Design, NoSQL Database
Machine Learning I - Machine Learning, Association Mining, Cluster Analysis, Recommender Systems
Visualization of Information - Data Visualization, Creative Coding, Visual and Information Design, Programming
Quantifying the World - Data Wrangling, Accessing APIs, Data Collection Design and Implementation, Synthesize Concepts in a Capstone Project
Plus, electives in Cloud Computing, Machine Learning II, Statistical Sampling, Natural Language Processing, Business Intelligence, Time Series Analysis with R, Data and Network Security.
Re:Analytics? (Score:5, Interesting)
It is indeed not comparable to a statistician. The core problem is getting the statistics and their semantics right. Getting the data is trivial in comparison. But these people will be used to replace statisticians and that will not end well.
Re:Analytics? (Score:4, Informative)
But some people apparently think that this is a good idea instead of getting a real statistician.
I'm not sure what program you're speaking of, but the data science courses I've seen pretty much include statistician coursework. A standard bachelors in data science is four credit hours short of mathematics minor and a masters in it includes at least twelve hours 500-600 level statistics. A masters could easily be used to work towards a mathematics bachelors.
Data scientist is pretty much a statistician that can do programming in at least two HLLs (Python, Julia, R, etc) and at least programming in C, C++, Pascal, Fortran. Understand data structures, sorting algorithms, uses of recursion, and so on. Basically, data scientist require the ability to pass level 300 comprehension computer science. Which depending on where you go to, includes assembly, but stops short of things like intro to OS concepts and the like.
In the mathematics, it basically includes all 100-200 level so a full two year in calculus and linear algebra. Where it changes is that is focuses on statistical 300-400 levels and skips on analytic geometry and post Calculus 200 level courses. Additionally, there's the 300-400 levels specific to the course that focus on SQL (that's DDL, DML, etc plus PLSQL/TSQL req electives), database concepts and management, project management coursework, and so on.
I don't know where you get the idea that a "real statistician" cannot be found in a data scientist. There might be gaps in some mathematical theory for the data scientist, but their ability to command a fundamental understanding of the foundations of statistical work, is one of the main focuses of a good data science program.
Re: (Score:3)
I don't know where you get the idea that a "real statistician" cannot be found in a data scientist.
Simple: I know one and I know how extremely hard it is becoming a good statistician. I also understand that statistics is an area where there is only the option to be good or very bad, you cannot really be mediocre. Hence I do understand what you do not, namely that basically none of these qualifies for what they are supposedly able to do. They will do a lot of damage though.
Re: (Score:2)
Re: (Score:2)
I am not a statistician. Not even remotely.
Re:Analytics? (Score:4, Insightful)
There is a really big difference between statistician and data science. It's about the same as as the difference between math and engineering. Just because one requires some knowledge of the other doesn't mean you want a mathematician to build you a bridge.
Re: (Score:2)
I do however want that engineer to use solid math and actually really understand what he/she is doing. That quality seems to be absent from "Data Science".
Re: (Score:2)
Not necessarily, computers are for a whole lot of things. Setting up an e-shop, enabling 2 ways video chats, making sure you car brakes on time, ...
Data analysis is only one thing that computers are used for nowadays...
Most data analysis is not automatized. You'd be surprised how much analysis is done by dumping stuff to CSV and giving that to an "MS excel guru".
What's the point when truths are "offensive"? (Score:1)
What's the point of doing this so-called "data science" when the objective findings are so often deemed to be "offensive" and discarded just because they don't correspond with narratives being pushed by academia, the media, left wing politicians, and other left wing organizations?
Crime stats are one particularly sensitive area, for example. A closely related area is those matters involving illegal aliens. Climate related matters are yet another. The raw data and objective analyses of these data end up highl
Re: (Score:1)
In a lot of places, the main task of the "data scientist" is going to be making sure the executive's "dashboard" program (which gives a dumbed-down "overview" of supposedly relevant data) uses the correct combination of pretty colors.
Re: (Score:2)
Re: (Score:2)
> 2012
> Somehow represents all 50 states and millions of people worth of the Republican party in 2019
Somehow I don't think you've thought this through. This is the weakest weak argument I've seen in a while.
Bubble about to crash ... (Score:2)
Remember when everyone with half a brain flocked to universities for technology degrees when the internet was new in the second half of the 1990s?
It ended up with the Dot Com Bubble [wikipedia.org] in 2000.
We are due for a recession (or massive correction), since the last one was 11 years ago.
This article is just indicator for it ...
Re: (Score:1)
Remember when everyone with half a brain flocked to universities for technology degrees when the internet was new in the second half of the 1990s?
Don't you know? The job market is like goatse's backside - it always expands to accommodate everyone with a college degree seeking a well paying job! You can't have a glut of labor, because labor doesn't follow the laws of supply and demand, the right wing folks told me so!
Re: (Score:2)
Data Scientists and more general Analytics resources are able to find ways to drive cost savings in ways which other areas might not be as well suited to drive.
In addition, data clearly holds monetary value; however, you need individuals within your organization who are adept at finding, using, and distributing this information to others in a way which an organization may capitalize upon.
We haven't even begun to scratch the surface of meeting the demands of industry with our available Data Science talent po
Re: (Score:2)
So just a buzzword for MBA, got it. I suppose it makes sense for Julia, R and Python to replace good old Excel+VBA, as the latter are increasingly inadequate when dealing with large (dimensions,volume) corpus.
Re: (Score:2)
We're arguably still in that first recession ...
A professional degree to invade privacy. (Score:2)
Re: (Score:2)
Indeed. These students are mostly taught to look for correlations in big datasets using algorithms. Cambridge Analytica, for example, used such correlations to create psychological profiles, which they then sold as 'likelyhoods' that someone thought a certain way. It's basically what the entire databroker industry does: selling correlation as causation.
Yoshua Bengio, the 'inventor' of deep learning, worries about this a lot. He likens modern 'data science' to alchemy, and doesn't believe calling it science
Important job training (Score:2)
Learn the phrase "do you want fries with that?"
All with +10 years experience in AI, IoT and ... (Score:4, Funny)
Correction! (Score:2)
They will produce around 2,000 Analytics Masters.
The other 8,000 will be graduates with a diploma that claims they are masters.
https://qz.com/967554/the-five... [qz.com]
You can go to college to receive an education, but it does not mean that you were educated!
I remember one of these from when I was (Score:5, Insightful)
in my freshman year. Statistics teacher, PhD, relatively young, very self-confident and depressingly knowledgeable about the intricate details of using various software packages to churn out various numbers.
So, third week through the course, and I'm trying to derive a useful statistic or two for an experimental study of one particularly absurd system with weird energy distributions, and I am tired and overworked and hit a small snag in a simple geometric probability calculation (it was a volume integral) and I go to ask for advice. The guy looks my calculations over with a blank stare and says... "I'm a statistician, I don't know any calculus". Luckily, I still had time to change the course.
I bet a lot of the "data scientists" are like that guy.
Re: (Score:1)
I'm a data/computer scientist. I don't even remember the last time I bothered with math. awk oneliners pretty much solve everything
Re: (Score:1)
Why do you need a four year degree then? I've read the awk book, it's not close to even a semester worth of material.
Re: (Score:1)
You're not a scientist, you're at most a data cleanup hygienist.
Re: (Score:2)
That's a funny story, but I presume he was pulling your leg. I took quite a few graduate statistics courses (I'm an EE) and they were brutal (and extremely calculus-based). In some cases a single homework problem would be literally pages upon pages of calculus. It sucked. For example, calculating the effect of a given linear system on the noise distribution of a random signal is pure calculus.
At the higher levels, Statistics is as calculus-based as engineering.
Re: (Score:2)
You misunderestimate the number of math-illiterate engineers and statisticians. And the number of worthless PhD programs.
market demand (Score:3)
this is created by demand, you read nothing else (even on /.) that data is the new oil and that there aren't enough data scientists to meet demand, etc.
obviously a lot of people might think it's a good idea to get some sort of degree that enables them to enter this market.
Similar to Computer Science (Score:3)
There aren't *that* many jobs (Score:2)