Anyone in data science or analytics? what do you do? what was

>>1081678
OP here (sorry for the ID change.) I work at the NSA and use data mining and social engineering techniques to screen online forums for suspected dissidents.

Analytics, mostly SQL/R/python

Self taught, I came from a science based degree, found it quite easy to pick up, been doing it 3 years now.

Live in the North UK, so pay not huge, about £45k a year or so ($65kish), work for an IT consultancy.

Im considering Data Science, been studying R/Python and i come from an economics background, but i cant decide what will be more profitable, pursuing a masters on data science or one in finance/CFA. From what i understand you could very well apply data science to finance, specially machine learning and so on right?

Yeah, but its an incredibly complex field, and you often need experience already to get into it.

Often better cash is found in the consultancy/startup space, and its a lot nicer to work in than the banks also.

In terms of skills, machine learning is very hot at the moment, but as you can imagine, very complex. Emerging tools also may replace the need for it, the fields is moving very fast right now.

Honestly though, SQL/R/Python - can't go wrong!

>>1081738
>dissidents

In burger land those are called domestic terrorists comrade.

Data Scientist at a large well known technology firm here. I work on marketing projects using data mining to calculate churn rates and how to more effectively spending marketing funding.

I have a PhD in Economics. On my team we have a PhD in Computer Science from CMU and a PhD in Physics from Stanford. Looking around the industry with people I know, it would seem that the common expectation for a data scientist position that is just not an advanced excel user the expectation is at least a masters in a quantitative field for education.

I started about a year ago and make around 95k but I am in a mid-tier city on the east coast. Personally, I think the salaries for data science positions are overblown because for the high salary positions the expectation is normally a Phd and multiple years of experience (10-15 in machine learning). Pay is still higher than the average university teaching gig however.

We use R/Python/SQL for programming languages. Some groups are using C as well but that may be just a holdover from previous work. We also are using Spark as another technology. Some people I know in other industries use SAS.

Feel free to ask me anything. My personal feeling is that to break into the field is to get a masters in analytics or data science as they spring up at well respected schools. North Carolina State has a great program that has superb placement.

>>1081764
From my understanding it is easier to go from quantitative finance to data science than the other way around. However, the people I know who work as quants work 20-40 hours more a week than I do as a data scientist. I personally don't think working twice as many hours a week is worth an extra 10-30k a year.

>>1082134

What do libraries like TensorFlow and Torch actually do? I get that they use arrays, but that doesn't give me much of an explanation.

>>1082134

Is

>https://www.quantopian.com/home

legit, or is it a scam?

>>1082134

If I wanted to run my own home DIY data analytics, what would you recommend I start with?

>>1082134


If I wanted to make a DIY Watson to answer questions for me, where would I start?

BS in econ.

C# ( worthless)

most of what i do is sql, excel minitab.

Learning R, python and hadoop

I work for a major manufacturing company that produces household chemicals.

I am studying for my CFA 2, and my company is paying for my greenbelt.


2017 i start my masters in financial econ, on the company dime.

>>1082435
Quantopian is legit but fucking good luck if you think you can make an algorithm that beats the market on a risk-adjusted basis. If you win that monthly competition you'll probably make decent money.

>>1082436
I have never heard of them but there is a swarm of companies peddling data science solutions. As a disclaimer I don't work in the financial sector.

>>1082436
My question is what are you looking to accomplish with you home made analytic engine? I know that the company I work for has made a multiple analytic tools with many not reaching significant usage because they fail to adapt the tools for use with popular solutions such as python or r. A lot of the tools have already been created to be honest unless you have something truly novel in mind. I am amazed at how fast solutions are generated that replace a lot of specialized work.

>>1081678
Current analyst/scientist. Mechanical engineering degree, self taught sql, python, c#. Currently at 70k with options.

>>1081678
I come from a finance background and I'm looking to learn how to backtest and develop algo trading strategies. Currently teaching myself Python with Codeacademy and others... anything you guys would recommend for a beginner?

>>1082594
how useful do you find c#?

I have not used it at all , and have let it slip.

>>1082629
R

if your also doing python it is a no brainier combo

>>1082634
I had never heard of R until I started looking into languages to use for data analytics. Sounds like it is going to be in demand

>>1082134
I'm an econ major that took some cs courses on the side. I'm taking a course on data mining atm since I thought it might be useful. So far, it's pretty interesting.

Considering the difficulty of data mining and data science why are the salaries so low compared to software engineering or development jobs? Also seems to require a lot more knowledge and educational background.

Also, is this a good gateway to graduate school? I love economics, but I'm not a mathy kind of person so I struggle with the more quantitative stuff (I'm good at cs at least). Also, how do you break into the industry?

>>1081678
I work in Analytics at a startup in the southeast. BS in CS from a top 5 Computer Science school, 90k starting + options. Interned at the analytics team at a consultancy group the summer prior.

I work primarily in Python with a constellation of other tools, notably Elasticsearch and SQL. I've been learning Pandas as an alternative / supplement to R.

>>1082631
I use it regularly. Mostly in support of ETL work and scripting. We us SQL server and SSIS so c# is native. R and python are also very helpful. Putting those on your resume will get you offers like crazy.

>>1081764
if u canada, take a look at UofT MFE

>>1082644
It is the language of data analytics.

I would also suggest learning how to use Periscope for dash boards.

Dashboards are easy to make but c suite eats it up.

>>1082647
The trade off is a much more stable job.