Redpill me on "data scientists". Some of them work

>>8445960

They're just statisticians with a meme name.

>b-but muh neural networks

Statisticians.

>>8445960
As your article suggests, "data science" is basically the business buzzword of the decade, so depending on your company you may find "data scientists" working on everything from data acquisition to data preprocessing, data warehousing, data cleaning, data analysis, data visualization, forecasting and even data-driven (i.e. evidence-based) policy recommendation.
Don't expect HR to know the difference.

The more tasks a "data scientist" handles simultaneously, the more indispensable they become to the company (especially if it's a small company), which might explain some of their smugness.
That said, it's statistically more likely that the people you meet were assholes.

t. data scientist

>>8445965
Get given data. Put in in R. Coffee. Run a few tests and browse the web. Lunch. Run a few more tests. Coffee. Print a boring graph/write a report. Home.

>>8445970
This description is spot on. 100% serious. I know an older guy that got a senior position at a tier 1 firm doing exactly this all day and he was completely illiterate with computers (e.g. wouldn't even know how to reinstall an OS) - they taught him everything.

£5-6k/month net.....

>>8445960
i worked as a data scientst at cambridge get on my level losers

also it's mostly ETL and hten some modelling i dont know if people keep saying it's stats cause it's not really. it's more techie and less p values. i'm pretty shit at stats, but i'm kaggle master and do well in data science. stats is a lot more mathsy imo

>>8446037

It's still statistics, you're just outsourcing a lot of the logic to programs instead of doing it yourself the way older era statisticians would have to.

http://www.merriam-webster.com/dictionary/statistics

>a branch of mathematics dealing with the collection, analysis, interpretation, and presentation of masses of numerical data

>>8445965

At least some of them has real math background. Problem is some cs grads try to pretend they know stats and do "data science".

File: 1477834696979.png (150KB, 848x1200px) Image search: [Google]

150KB, 848x1200px

Download R-Studio. Go on stack overflow/favourite help site when you don't know how to do anything. Remember to drink coffee so you look busy. Can confirm, this is my job.

>>8446044
i guess, but i think a new name like data science helps. statisticians use excel and data scientists are trendy programming rejects rolling jupyter notebook with python and ggplot. i wouldn't stumble upon my jobs if they were listed statistician. and then in interviews i just talk about my cray machine learning skills and look at my neural network implementations and natural language processing. statistics is overloaded

File: punchingfrogs.gif (483KB, 785x757px)

483KB, 785x757px

>>8446046

>listing "database" as a skill

>>8446049
it says SQL and noSQL though? how is that not a skill

>>8446046
Someone please explain this chart to me?

Being good at everything on that sheet is pretty much impossible unless you have no life.

The math and statistics section is a 4 year degree.
Programming and database is another 4 year degree.
The bottom two sections together are probably another 4 year degree. Maybe a 2 year degree from a decent college.

If you're not good at something on that sheet, you need to defer to someone who is. That's why we hire multiple people to work together. Sadly these dumb fucks keep cramming more and more skills onto checklists until we're spending 20 years to be qualified for a job that pays 75k. They can't drive the price of tech workers down, so they increase the skill requirements.

>>8446056
i'm good at all of them get wrekt brainlet

>>8446056
I don't know what kind of shitty school you go to, but they're all covered in one CS degree.

>>8446058
t. typical arrogant data "scientist" dickhead who wouldn't even know how to reinstall an OS

Can someone explain to me why is everyone here saying shit like "data scientists are incompetent / computer illiterate / know zero programming languages / literally do nothing" and similar stuff, so often?

Is there maybe some truth in that or is /scv/ just being jealous?

>>8446063
bitch i've been running arch linux on my machine for 3 years with no kernel panic

>>8446056
Not really man, if you do a dual CS-statistics degree you should have no problem mastering all of them, the bottom ones are just personal skills and personality attributes.

I am trying to get into the big data field and am currently doing a masters degree in survey statistics. I got no background in computer science or anything like that. The only stuff I know is R programming. What else should I learn?

>>8446076
cs grads will always be chosen over you

>>8446077
I am only 22 years old so I could easily do a bachelors degree in cs afterwards. The question is, is it worth it?

>>8446082
no, but you could do a 1 year masters in data science at a top university instead of survey stats lel

File: Data Science (2).jpg (967KB, 620x2837px) Image search: [Google]

967KB, 620x2837px

>>8446077
(not true btw)

>>8446080
so much bs

>>8446083
Don't have the requirements for a masters degree in data science, you need a background in cs stuff for that where I live

>>8446050

>how is SQL not a skill

Is the bar really this low now? SQL as a "skill" is one step above Microsoft Word as a "skill."

>>8446085
You sound insecure.

>>8446090
SQL is a skill out of 100. making efficient SQL queries without joining too many tables, good database design, etc are valuable skills and DBAs make a living from this. also the SQL was grouped with NoSQL which already implies lots of stuff on top

>>8446092
the one who wrote that shit is the epitome of insecurity lmao

File: GenericWhiteWojak.png (126KB, 619x757px) Image search: [Google]

126KB, 619x757px

>>8446093

>making efficient SQL queries without joining too many tables

How does one become a data scientist?

>>8446099

Apparently you just have to hold literally any office job for a couple months and then list SQL as a skill.

File: 1431956327975.png (598KB, 722x525px) Image search: [Google]

598KB, 722x525px

>>8446093
>SQL is a skill out of 100
Administrating an SQL database, possibly.
Not
>making efficient SQL queries
Take a look at the chart >>8446046 again, and note the skill that comes immediately after SQL in the list.

>>8446056
this.

They're probably just dumb CS graduates that use Excel's graphs and some silly statistics/machine learning tool to get a wrong result on some silly statistic like the ratio of males to females visiting the company's website.

>>8446106
sql is a superset of a relational algebra implementation, so knowing relational algebra is not enough to know SQL

File: smug.png (4KB, 149x136px)

4KB, 149x136px

>>8446122

Yep, that's what the "Story telling skills" is for under the communication and visualization header.

You don't have to produce anything actually meaningful, but you have to convey it in a way that is believable.

This is the current state of the world we live in.

Combine that with an attitude smugness that will make somebody attempting to question your logic feel uncomfortable and you got yourself a keeper.

I see it all the time.

>>8446037
This gives me hope. Right now I am in CS. I am only okay at stats. I could code monkey my way to a data """scientist""" job apparently.

File: 1350054576901.jpg (44KB, 500x258px) Image search: [Google]

44KB, 500x258px

>>8446093
>he fell for the not joining too many tables meme

File: frog_rising.jpg (23KB, 438x438px) Image search: [Google]

23KB, 438x438px

Google "backyard data science" and you should find the blog of a guy who writes quite a bit about it.

Data scientists are good at making infographs to make them look more attractive and useful, when really they are nothing more than glorified business grads

>>8446067
I could crash course someone with python numpy pandas and some bash with piping for "big data" data munging and get them to the "acceptable" level within a day or two. Its super basic bitch shit.

>>8445970
>R
But my datasets are 300+ gb

>>8446301
Which programm would even be able to manage this

Can you get a job as a data scientist if you graduated with an engineering BS?

>>8446255
Get on that, animepro.

>>8446305
Distributed computing. Dataset on hdfs, write a distributed map reduce program in whatever language you want (i like scale), and run it on your server cluster with the hadoop or spark platform

/sci/ conflates data science and business analytics. R is for small scale analytics, where data science is platform agnostic and requires a technical toolchain expertise, domain expertise, and a scientific methodology

>>8446474
*scala

File: image.jpg (46KB, 452x601px) Image search: [Google]

46KB, 452x601px

>>8446084

File: 786778678.jpg (24KB, 279x209px) Image search: [Google]

24KB, 279x209px

>>8445960
i think they do this

https://www.youtube.com/watch?v=yaCDHrW8aO8

please no jokes about the title

>>8446056
Brainless brainlet

>>8446080
These nu infographics really piss me off