/dsg/ - Data Science General

is Python not the cool data science language anymore?

why would you want to be a """data scientist""""
that shit is boring

You make more money than a typical country's (including the US's) president (bribes excluded) and can work anywhere you want, for any type of company, and often even from home.

I'm actually using Apache Spark and Scala.

>often even from home
They encourage work from home because office space is expensive.

who's "we"?
>tfw even AI gf starts out with chad
what are the course's prereqs?

this but MFE and unironically
where are these 6-figure jobs? how much do consultants make?

Anywhere in north america. In toronto or montreal you make 120-125k MINIMUM starting (even at startups), consultants make 100k minimum. In sv you're closer to the 7 fig mark. It's ridiculous, it's a massive bubble, and it won't blow until 2 decades from now. But when it does, the aftermath will be hilarious.

Attached: nope.webm (480x374, 40K)

Web dev codemonkey here, what are the best free online resources to learn data science shit?

>AI need micro transactions

PRE-REQUISITES

Linear algebra (vector and Euclidean spaces),
differential calculus (Jacobian, Hessian, chain rule),
Python programming,
basics in probabilities and statistics (discrete and continuous distributions, law of large numbers, conditional probabilities, Bayes, PCA),
basics in optimization (notion of minima, gradient descent),
basics in algorithmic (computational costs),
basics in signal processing (Fourier transform, wavelets).

Making shinyapp, i need to have every variable in script, otherwise app dont work on remote server. How to remove redundancy? Should i learn functions?
Code in next file

server.R

library(shiny)
library(jsonlite)
library(curl)
vectorofCallsNamedPrice2

I can use Excel. Does that make me a data scientist?

>Data Science

Attached: 0cbpdjii3pa11.png (645x729, 105K)

To be more like Nate Silver, and to know how to efficiently scrape and archive automatically in order to spend less time searching around online

which is it then

RapidMiner > R

I work exclusively with SQL, VBA, Excel and some .Net.
What else do i need to enter data science amd make bank?

>I know precisely nothing
>what else do I need?
lol

>>I know precisely nothing
I mean my current title is Data Analyst. Its pretty much a jr role. Idk how you could say I know nothing when i already listed what i know.

Aspiring computational linguistics masters here, what am in for bruv

Attached: FE228B50-7727-4372-9108-CA17EF73253C.png (450x222, 27K)

bump

I have a bot that trades with other users Steam Trading Cards.
Currently hoarding as much data as possible which later will get thrown into Tensorflow.

Not much, its one of the easiest masters and the one that is least likely to prepare you in any shape or form for any real application. Sorry senpai.

Source: Actual researcher. For a university. One that Amazon and Google recruit from.

Still less than 3dpd gf

I've got a summer internship as a data science intern, learning python as that is the language of use

Any other advice

read the thread lmao

smile and agree to everything your superiors say.

PCA isn't under probability. It's under numerical linear algebra.
Nice starting list tho, should be stickied so that people stop asking what they need to know.

Data analyst is the new name for data entry specialist, which is the new name for secretary. You are not on a data science path at all.

This. Especially if they ask you to make a turing-test-breaking bot within 2 months. Just smile and agree. Don't try to make them understand how retarded they're being. Just smile and agree, and every time you have the opportunity, point out that your manager is being niggers even if that's not true. This will ensure the failure to accomplish the impossible task will be credited to your manager and not you and might net you a good position in the company after the internship.

Thanks for your input.

Honestly I come from a poli sci background (I have an associates in Arabic and actually enjoy linguistics, which is half of why I want to do this), so maybe that’s not such a bad thing. I’m learning python at the moment and I am trying to figure out how to make limited dictionaries, I don’t want to be pajeet level with programming.

If the workforce wouldn’t be too friendly as I finished my degree, perhaps I can at least gun for a PhD; I had formal linguistics, neuroscience/cognitive science, further down the compsci rabbit hole, or maybe even something like bioinformatics in mind, that seems to be the beauty of the major.

New fast.ai course is coming out in January, works with pytorch 1.0

Your expectations are so completely off it hurts to read. Go talk to people in the know, both in and out of linguistics, and see what they say since you shouldn't trust people on an anonymous image board anyway. The fact is though, linguistics is downright deprecated and has 0 applications whatsoever nowadays if you're trying to be a special snowflake variant of code crackers (others would have instead went through stats and cs which is a lot more relevant).

Wouldn’t knowledge of NLP and machine learning be useful on your CV? I figure that one’s usefulness also depends on their own portfolio in addition to their degrees anyway. I truly enjoy linguistics, so that’s why I’d like to choose this field

What do you do for research? What is your direct experience with computational linguists?

I'm stating a physics degree at the beginning of the year, and If I am capable enough want go the mathematical/theoretical path. The units involved in the mathematical course involve data science and supercomputing shit.
To get an advantage since the pass rates for some of these classes is 20% since the majority has no past experience with the programming languages involved, could you guys list or recommend some of the easiest or best languages for that purpose.

Linguistics is not ML and has nothing to do with ML, that's your first problem it seems: you don't really know what these terms mean/how they relate. Classical NLP has been dead for years as well and is different from linguistics anyway.
If you were to focus on DL you would have the flexibility you're aiming for but you wouldn't be focusing on natural language in particular (though you could go with a prof that particularly likes natural language and work on that during your master's, remembering all the while that the primary focus is not natural language but DL).
Since DL (and ML in general) is so hot right now, having it in your resume allows you some fantastic job opportunities, money-wise, but also ensures you'll be a slave until you get a PhD and basically be in charge of annotating large datasets by hand because HR is being retarded and companies are looking for such slaves. Of course, you would be barred from even that without a master's. That's how dumb they're being.
I have a master's in DL and I'm a PhD candidate in bioinformatics. Previously, I worked at a startup, and then a big-5, on dialogue systems and NLU. Given that part of the job description is to keep up with all the developments, in and out this sphere, in NLP, I know quite a lot about the topic, even beside my actual experience. For example, I can tell you that every company ever tried to integrate classical computational linguistics in their DL-first teams and all of them failed every time. They're not exactly hot about the prospect of trying again.

That doesnt apply to any of my previous posts. Secretarys dont write code.

Thanks for the insight, and I have a stupid question:

ML aside, is NLP primarily statistics-based as it seems to be from the papers I've read, or does it use linguistic and semantic concepts to help translate?

>python
dropped