Data science

That's help desk

Because you ought to be making a thread and informing yourself before posting opinions you don't know about? God it's annoying when a fucking zoomer enters the thread

Pay attention, loser. You made a thread about buzz-word that doesn't mean anything and provided an infograph that is full of nothing but other buzz-words. There's literally nothing to add to it. There is no runway on this thread. You want to know about data science? There's nothing to learn about data science. It's a term invented to try and leverage normie ignorance and buzz about things they don't understand into something that seems sophisticated, academic, and legitimate.

Pseudo position meant to show the figures management wants to see. Usually the people who get in know someone who knows someone. Try applying for one with a fake profile based on a real person. You won't even get a call back. That's how I know it's rigged.

Every zoomer wants to be a data scientist these days. Not sure who planted this idea into their heads.
These Bayesian analysts should stay away from anything that requires high precision engineering though.

got my MS in a less tech-oriented field just to get swiped for a well paying job to learn SQL. These Data scientists aren't all that great. Nice enough folk though.

Kinda just biding my time until I can get back into public policy with the new "programming" skills I pick up.

this. everyone I know got their job because a relative works there.

elaborate what? it is what it is, bullshit.

Fake job.

absolute s~o~y shit

The amount of illiteracy ITT
Yeah code monkeys are the real intellectuals.

My sister is a data science at the VA, and I know some R. that's it.

based and redpilled.

Took a few classes, then snagged an internship. It's 50% sanitization, 25% googling R/SAS functions, and 25% writing reports. The "storytelling skills" thing isn't a meme — I literally write my reports like YA short stories, and the managers praise me for it.

Best part about the job is how self-directed it is. I report to a woman with no math/technical skills, so I can tell her "this will take a week" and she'll take my word for it. I hammer out the report in 4-6 hours, and I'm free until the next Monday.

Worst part about the job is tagging along to meetings just so I can read the reports I've already written. Apparently MBAs are illiterate.

t. English-major-turned-wageslave

Attached: all_gay_here.png (500x500, 283K)

>ah yes data my most esteemed interest I go home, pour a gin and tonic and read database rows all night

chortled audibly

What classes qualified you to get an internship? It's interesting you mention the English major because the skills for statistics and English actually go hand in hand. t. Have high verbal IQ and reasonable quantitative IQ and enjoyed statistics

Pretty cool, user. Sounds comfy desu. Do you ever get bored? What do you do to keep yourself occupied?

Cobbled together a CS minor with:
>Intro to stats
>Calc 1 & 2
>Intro to statistical software (taught in SAS, but the skills transfer over to R)
>Java 1 & 2
>SQL 1 & 2

I VPN into my home computer and write on company time. Hoping to finish revising my debut novel by the end of the summer, so I can see if I'll make it.

A bit of advice for future wagies: pay attention to gossip. If you save someone from a disaster in their personal life, they'll put in a good word for you at the office.

bunch of frauds, software engineering team has to constantly hold their hands and fix their shit

data basedence

Lol just graduated with a MSc in Applied Math and already got a job offer and a few interviews (most of them wanted more of a developer than what I wanted to do so I turned then down). Not that hard.

The industry is all about lying with statistics. After 5 years of a PhD i realised that my supervisor wasn't going to give me a PhD (i.e. recommend to everyone my PhD sucked) because i discovered a bug in his analysis software that made his and 100 other people's analysis incorrect. So he cut my funding and i left with nothing.

Made €1300 a month for 5 years for the pleasure.

Datascience isn't a real thing and already we're seeing the public catch on. Take this blackhole bullshit for example. ohwel

These pics are so fucking cringy holy shit.

Where did you do your PhD?

consulting guy here

>get """data""" from customers
>usually bad encoding and/or weird file types like huge excel files because "that's how our database exports it lol"
>usually customers have crap data, otherwise they would have already done it
>spend two weeks sorting shit out
>eventually figure out they send the wrong data
>customer denies, then send the next data blob
>user, WE NEED THIS NEXT WEEK
>power mode on
>throw away all data that is annoying instead of doing proper data preparation
>label all weird data as "weird" so I can report "10% of the data is weird, we ignored that" later on
>finally get a data basis and pour it all into AWS because fuck the costs
>rape data with ugly queries with 8 levels of subqueries because fuck normalization
>copy and paste old powerpoint, change charts and subtitles where necessary
>"whoa, just in time user. good work! but the way the meeting is canceled anyway, so we need this in about 2 or 3 months"
>knowing work will eventually end up in a single slides on a meeting
>knowing MGMT will ignore the data evidence anyway
>charge them the AWS bill and call it a day
>so_this_is_what_love_feels_like.jpg2

>hacker mindset
What the fuck does that mean?

Max Planck Institute

i feel this in my soul.

Basically someone who finds creative and unorthodox solutions fast.
>catb.org/~esr/jargon/html/H/hacker.html

But here it's only a marketing term to attract the right people. Who would apply if they wrote something like "has to bear with unpatient management, so you better bring good results quick OR ELSE" ..

Attached: what we say - what we mean.png (599x750, 350K)

more like "willing to work for free"

Man I'm sorry to hear that. Also can't you sue for wrongful termination or something?

According to that image I'm data science meme at work/school(bizstat), its easy after everything is setup. Nobody calls it data science but analysis/stats/bizstat.
>Mentions what looks like a javascript library.
I don't do web stuff but its odd javascript is even mentioned. I don't use python at work at all, I think python is fine though. R is better.
>No mention of excel or crystal reports.
That's weird, excel should be mentioned. Is everything mentioned open source stuff, its another shill isn't it.

based and schizopilled

Hackers are criminals unless they have a tv show. So in relation pushing a certain mindset using statistics basically, lying like the big boys.

I know a couple of people from there, this isn't the first time I hear about dishonest supervisors there.

>i realised that my supervisor wasn't going to give me a PhD (i.e. recommend to everyone my PhD sucked)
Mind elaborating? What did he tell you?

basd

I work as a """data engineer""". Thought it was pretty cool since you got to work on a lot of meme technology instead of one or two things like your typical software engineer. You don't even need to be that good of a programmer anyway since most things are done in batch, who cares if it runs in 15 minutes instead of 10. Also you're pretty much indispensable since data scientists are mostly technologically illiterate and they suck your cocks over trivial things.

I feel my soul crushing on your behalf

you are not technically employed, but more importantly, when the question is whether or not an academic decision is "just" or not is always the universities prerogative. I had a similar experience at Uni for my bachelors, where i did this project that was worth like 30% of the entire grade. It's important if you want a good grad to do well. So important is this mark, that two separate professors mark it and you get an average.

In my case, one prof gave me 80% and one 20%. I'm not being hyperbolic or rounding those are the exact marks. They gave me 50% which is like a shit-tier pass. I complained and said the guy who gave me 20% is just confused or something, I spent months on this - but despite escalating all the way to court, i was told the institution has the right to grade however they like, it's not up to the courts to determine what is or isn't good work.

So tl;dr no there's no option for stuff like that. Academia has a monopoly on truth.

yeah this is just me i know plenty more shady shit. Professors who only ever hire outside Data Scientists to do the analysis of their data so they can maintain plausible deniability if the statistical fuckery is ever detected - all the while everyone knows the professors tell these ad hoc workers which programs to use, how to use them, which result they need to produce, etc. Shady shit man. And plenty more that would de-user me.

I was writing at home because if i worked in the lab i'd be called to continue doing experiments (supervisors see any time spent writing a thesis no one will read as wasted time, and they're not exactly wrong...) but he couldn't code and the whole project was about data analysis. So we had nothing to say, i got no support, I received an e-mail i wasn't going to get my €1300 next month and i should be prepared to defend my thesis in a week - wasn't ready, couldn't pay my rent, started sleeping on my friend's couch, went mad, the rest is history...

>hacker mindset
You can't be serious

> Not really a programmer
> Not really a DBA
> Not really a Accountant
> Not really a Business Manager
Basically a Master of None

>tfw getting my masters in this after a background in Software Engineering / Programming
AMA

>data scientist
buzzword meme used to refer to people who can write code to query a database and do some simple mathematics

data """science""" is the biggest tech scam of all time

Funny way to spell Unix.

story time
>company I work for fired 7 (((data scientists))) about a year ago
>I was made responsible for getting their shit working again
>all of their code was stored on Google Drive, no git tracking
>code quality was like something you'd see in one of those retarded CS grad memes
>one part literally looped and incremented a variable (presumably because the author didn't know how to fucking multiply in Python)
>another script took 10 minutes to query a 1,000 row database and find what country had the most sales
companies don't need (((data scientists))), they need programmers with decent scripting skills who actually know how to write maintainable code

How does it feel to be scammed?

Easy bachelor in statistics and master in CS.

Precisely this.

Also

>influence without authority
enjoy your eternal zero respect

I know that feel. I did a master's in """networking""" which literally consisted of learning a couple meme frameworks made by the professors (which don't even work well) and pretty much nothing else.
Having it on my cv did help with the job hunt, but it's not an exaggeration to say that I learned literally nothing in school during those two years. At least I was on a generous scholarship, so I spent very little and I was never in debt.
I'm thinking about going for a PhD in the near future, but this time I'll take my time beforehand to make sure I don't end up in a meme program once again.

>21th
Ok retard

90% of my tuition is paid for through scholarship so I can't really complain. No loans and I'll get offers out the wazoo for going to a top tier school even for a meme program.

Also thinking about a PhD after a few years of industry. On the plus side, the program did get me interested in machine learning.

>t. seething underperforming brainlet
Nigger, if you were retarded enough to go for a stipend funded doctorate, it‘s obvious you shouldn‘t be in academia at all.

oh fuck i got all except that :(

Attached: file.png (247x41, 7K)

fuck are you talking about most people have to pay.

Not in first world.

>that entire picture
this is why the real tech-industry is dying and the marketing industry is booming. silicon valley is slowly replacing innovation with algorithms and technicalcompetence with social engineering. this is why we live in a world where you're considered a hacker for being able to change your homescreen to a custom image

Had a university internship in "big data" for 2 years and I had LinkedIn people approach me for "big data" jobs (even Amazon). Rejected them so I can work for the decentralized web because I don't give a fuck about big data and don't care where it goes.

>21th century

So a computer guy that sucks corporate cock.

Fucking hell, will the stupid buzzwords in modern business that are done to keep out people who they judge "unworthy" ever end?

I went from US to UK to Germany. I'm at the top of my shit. I looked at hundreds of potential labs before choosing another internationally prestigious lab to do a huge 4 year project in. I've seen it all. You either get a tiny salary, you gotta teach (US), or you pay. Most people pay. Salaried positions only take the best. Every year at my shitty institute we had 600 people apply for 10 spaces, twice a year. I got in and realised actually it's just like Peter Thiel says - everyone outside wants to get in, everyone inside wants to get out. These degrees aren't worth shit it's all about cheap labour, or better, easy cash $$$

stats was better before data science bullshit

Most of Jow Forums are fucking brainlets on par with """data scientist""" who they criticize.

Industry need people who can write halfass decent script, sql query and run statistical analysis. Vast majority of statistician cannot program for shit.

Meanwhile vast majority of programmer will have uncomfortable stare when you ask them very basic questions like "What the fuck p-value even is?"

or

"Tell me again, what is expected value?"

So yeah, here it is, retarded mixture of both

>one part literally looped and incremented a variable (presumably because the author didn't know how to fucking multiply in Python)

Can some one explain what is wrong with this?

Looping with a counter is pretty basic stuff we do all the time in our data classes.

I don't even get paid.

data science has many different meanings. It could mean you know SQL, it could mean you have a Ph.D in statistics, it could mean you have a Ph.D in CS specializing in ML. It is basically just a business buzzword and these people get hired to analyze data. Can't say I would ever want the job. Imagine telling business MBAs valuable insights from data all day and they decide against it because they think they know better.

We get it, you had a statics class today.

Negro, a standard doctorand occupancy as a scientific assistant in Germany pays like €3,5-4k at unis and €4-5k in free market. And you don‘t have to be a genius, just good enough for a PhD. Hence my remark about stipended positions which are usually filled with brainlets who shouldn‘t be there in the first place.

I work in the field and I'll tell you how it works.

>Your a manager
>You want the smartest people writing the reports you skim
>Instead you get reports from boomers with years of experience but no idea how the modern world works.
>You need to find a position where you can hire very smart people for big pay, but not fuck up your heiarchy of middle managers.
>Isn't that just what a consultant does?
>Yeah but they cost to much and I don't like how they are so independent
>What if I just make a position require that you know how to do some basic trivial coding and stats?
>This is perfect boomers hate computers and you can't do modern statistical work without computers.
>now I can hire an in-house consultant on the cheap and reward one of my pet MBAs by making him their subordinate
.
It's literally just consulting. Except you work for the company. You spend all your time finding ways to optimize shit the software engineers could have done years ago but never thought to do because they can't calculate profit margin without a terminal.

It's pretty fun as no one knows more than me about anything I do. I am sure the engineers make fun of my janky code, but they still have to implement whatever I feel like shitting out that month.

I get direct access to the upper management without the bullshit jokeying since I am technically not on the management track.

It pays less than software engineering, but it's probably the best job for a technical minded person that doesn't want to spend more than 50% of their time coding.

Only thing that sucks is that the data they give you is always always Garbage and you have to make up your numbers 70% of the time.

>Bayesian statistics is for hacks.
Whoah there, buddy, I was about to just casually browse through this read and maybe read something interesting, until I saw your post. What's wrong with Bayesian statistics? You don't like simplified statistics that is simultaneously more fundamental than that frequentist garbage you've grown used to?

But in all honesty, what do you think is wrong?

Attached: maxresdefault-5.jpg (1280x720, 70K)

He's not wrong. Bayesian stats is literally data dredging. Without expert analysis you'll have retard tier hypotheses suggested by the computer, especially if you are working with an absurdly small dataset, or you are retarded enough to make other general statistical mistakes. And many data scientists are total memes who get BTFO after one year passes and their models are completely broken because they didnt consider the actual value of the data they built their model on, ignore outside variables, and also forget to zoom the fuck out and look at shit over years instead of within the year. See Nate Silver in 2015.

t. Jow Forumsbrainlet

Probably a codemonkey that struggled with undergrad combinatorics.

Updating biases != building a falsifiable theoretical model. Also, depending on how tight your DS tinfoil hat is, you may even be retarded enough to continue using your faulty model after seeing direct evidence to the contrary of what it suggests, whereas a non retard would investigate what invalidated their prior model and consider what new variables could be at play instead of doubling down on being wrong and adjusting muh confidence interval. In some of the most egregious cases I've seen, some DS will actually go back to their original datasets and continuously bootstrap sample their limited data to see if they can make some adjusted crackpot theory instead of just looking at the new observations in the outside world which clearly are different from their sample data.

>ah yes data my most esteemed interest I go home, pour a gin and tonic and read database rows all night
my sides

This thread already seems good, and what you said sounds interesting. My friend who worked as a software engineer consultant at BoA, and he says everyone hated him (he's not a retard of socially autistic, either), so I wonder if consulting has anything to do with it.

Data science also gets scammed by cloud vendors who sell them all the latest tech because they think they are going to be google. I scaled a sqlite database to 1.5B rows once for the lulz and it actually worked. You don't need every single Apache product together for your tiny ass dataset.

First of all, I assume you mean priors when you say biases. Bayesian modelling simply exposes the biases/priors that all regular statistics relies on at a more advanced level. What do you think regularization is, if not a bias you are *forcing* on your model to make it comply with some previous notion of reality, that the naive model won't support?

L1 and L2 regularization (some of the common ones) both have direct counterparts in Bayesian priors, so it seems like you should adjust your critique to apply to all of statistics (bootstrapping is manipulable as well, as you point out), as a sin inherent to all statistical modelling.

Never took it, just seems like introductory advanced probability theory, right?

People usually hate consultants in every sphere. They‘re „ideas men“ payed out the wazoo, that swing around barely getting into particulars of business operation, make some grand plans/strategies and fuck off leaving the seething wagies to implement them. Pretty comfy if you don‘t mind traveling a lot.
t. worked in management and software consultancies

Any more useful tips? I just started an online data science program and I'm coming from a completely different field (paramedic).

I work at Buttercup Games, ask me anything

Attached: 1554667702236.png (680x680, 188K)