rooshvforum.network is a fully functional forum: you can search, register, post new threads etc...
Old accounts are inaccessible: register a new one, or recover it when possible. x


Career in statistics/data science?
#26

Career in statistics/data science?

Quote: (09-19-2014 08:56 AM)Travesty Wrote:  

^ Peregrine

Did a woman take over your keyboard?

Data scientists can make alot especially if you start your own consulting biz.

All big tech companies have a large need for them to track user behaviors.

I am sure there are many other Big Data realms esp. financial, medical, energy, telcomm etc... where these skills are extremely valuable.

I would love to know more about this. Any success stories for example?

A whore ain't nothing but a trick to a pimp. (Iceberg Slim)
Beauty is in the erection of the beholder. (duedue)
Grab your life by the pussy.
A better question to ask is "What EXACTLY do I want out of life and what EXACTLY am I doing to get EXACTLY that? If you can answer that question truthfully you will be the most Alpha motherfucker you will ever need to be. (PapayaTapper)
Reply
#27

Career in statistics/data science?

Also interested in learning more about this field of work and thanks to those who have contributed already? I don't understand yet why this is not suitable for a location independent lifestyle? Anyone care to elaborate? And also most post are from 2014-2015, has the field and software changed a lot in recent years? Big data is mentioned a lot as one of the big upcoming fields, I can imagine it is changing rapidly.
Reply
#28

Career in statistics/data science?

The data science part isn't the part that's incompatible with location independence. The technical tools develop rapidly. But the output always needs to be digested into pretty presentations and delivered to decision makers.

The guys making coin get paid not just for number crunching (which is the part that could be outsourced or done anywhere, assuming that the data isn't required to be maintained in a secure facility), but for the face time in front of senior leadership and explaining the findings, convincingly, in fluent English.

Data Sheet Maps | On Musical Chicks | Rep Point Changes | Au Pairs on a Boat
Captainstabbin: "girls get more attractive with your dick in their mouth. It's science."
Spaniard88: "The "believe anything" crew contributes: "She's probably a good girl, maybe she lost her virginity to someone with AIDS and only had sex once before you met her...give her a chance.""
Reply
#29

Career in statistics/data science?

Quote: (03-12-2018 12:14 PM)polar Wrote:  

The data science part isn't the part that's incompatible with location independence. The technical tools develop rapidly. But the output always needs to be digested into pretty presentations and delivered to decision makers.

The guys making coin get paid not just for number crunching (which is the part that could be outsourced or done anywhere, assuming that the data isn't required to be maintained in a secure facility), but for the face time in front of senior leadership and explaining the findings, convincingly, in fluent English.

polar nailed it. And here's the kicker: the latter is worth more.

Any H1B can build a statistical model; ELI5ing the results and methodology to various stakeholders is a much rarer skill. Not to knock the number crunchers: building a really good model is an equally rare skill, but it's all voodoo to most people anyway. And when something's difficult to understand, sales wins.

This isn't to say that you'll never be able to combine data science with location independence, but it's difficult. Data science is an extremely collaborative activity, which necessitates that people can have face to face meetings. And it's tough enough to sell a unit head on your model results when they hurt his bottom line, good luck doing it from a thousand miles away.
Reply
#30

Career in statistics/data science?

Quote: (09-19-2014 07:36 PM)Peregrine Wrote:  

Preface: I'm talking about quantitative analysis/modeling as it pertains to financial institutions.

Quote: (09-19-2014 03:22 PM)kongzi Wrote:  

Would you recommend quant/statistics/data science/numerical modeling as a better career than IT? Is it more lucrative than web development and other IT work?

If you're good at math and enjoy it, I would recommend it over IT. Math/stats skills certainly have a longer shelf life than IT skills.

Salaries range from 70 to 300k depending on rank. You could make more, but then you'd be in senior management rather than doing the actual analysis. Pay raises are mostly driven by promotions and job hopping. New jobs are easy to come by.

Hours are generally 40 to 60 a week.

Quote:Quote:

What are the people like?

Depends on the complexity of the data that you are working with. As the complexity increases, the people working on it tend towards shy/weird/awkward/borderline autistic.

Quote:Quote:

Since Americans like bad boys and jocks more than geeks, would you say the prestige is low for these careers? Is it any better in Europe or elsewhere abroad?

I would say it's neither plus or minus prestige.

Quote:Quote:

Are you required to spend alot of your free time working on personal projects and keeping up with technology news, like for IT careers?

You have to stay apprised of industry developments, but that's true of any career. If you're just looking to punch the clock, no.

Quote:Quote:

I doubled majored in applied math and another STEM field when I was an undergrad. If I get my Master's in a STEM field (not Statistics or CS) and have some programming skills, how can I get a job in those fields? how can I catch up with those who have degrees in MS Statistics to get a job? Do I need to learn R and SAS on my own and then use them to create personal projects, like programmers have to? Or am I better off studying algorithms and data structures?

Which STEM field is your Master's in? Math is preferable (for obvious reasons). Physics is second best.

Learn R, SQL, SAS, Matlab, and Excel (and Powerpoint, because you will usually present your data with it). If you're proficient with all/most of those, you should be fine. No personal projects necessary.

Quote:Quote:

Or if I don't like my current master's program, should I just look for data analyst/etc jobs now? Or the better ones require a Master's?

Master's required, sometimes even PhD. But that depends on the company. My area requires a Master's, PhD preferred.

Quote:Quote:

Is it true that there aren't many entry-level jobs in data science, and most job openings require a phD or lots of work experience? If so, am I better off getting an entry-level job as a quantitative programmer or statistician/data analyst?

There's grunt work that has to be done, which can be done by just about anyone who is good at Excel/R/SQL/Matlab/Access.

Quote:Quote:

If I want to live abroad, should I apply for quant/stats/data/etc jobs abroad, or work in the US first to build up experience?

Doesn't really matter from what I've seen. Data is data. I've seen people go both ways. If only US companies will take you, work there first and apply abroad later.

Quote:Quote:

Since I ultimately want to work as a freelancer/work for myself, is it easier to get into that if I first work in stats/data science as opposed to IT?

Remote work opportunities are more plentiful in IT. Your main shot at freelancing in quant work is becoming a consultant down the road, either for a firm or on your own.

Very solid advice. Where to look for problems for personal projects?
Also what book(s) do you suggest for statistics?

A whore ain't nothing but a trick to a pimp. (Iceberg Slim)
Beauty is in the erection of the beholder. (duedue)
Grab your life by the pussy.
A better question to ask is "What EXACTLY do I want out of life and what EXACTLY am I doing to get EXACTLY that? If you can answer that question truthfully you will be the most Alpha motherfucker you will ever need to be. (PapayaTapper)
Reply
#31

Career in statistics/data science?

Quote: (01-25-2015 03:33 PM)cibo Wrote:  

So I'm a data scientist and have worked for F500s and start ups.

Expanding on Agustin point. The data science field breaks down into:

Data analyst
Data Scientist
Data engineer

This creates a range between technical and business. Analysts have more business knowledge and have to do more client facing work while data engineers have to program more with less business facing. There's also a bit of difference in the tools between tech companies and normal corporations.

That said the common skills you need are:
SQL to get data from databases and is a fairly easy common language used for many of the Big data tools, Hiveql for hadoop, CQL for cassandra, etc. Most of the relational databases (mysql, teradata, orcale, sql server,etc.) are pretty much the same with some differences in functions and syntax here and there. The NoSQL databases have a lot of different types and can be a bit confusing if you don't understand relational databases enough. I'd work on those later after you have a solid grasp of Relational databases since the NoSQL's db's were created in response to the limitations of the traditional database.

Statistical programming language: SAS for corporate, R/python for technology/startups. This need somewhat depends the amount of raw statistical/machine learning work you're doing. Also when you start getting data from websites and apis, these start becoming more important.

Data visualization tool: Excel/VBA for coporate, d3.js for tech, and/or Tableau for any type of firm. You also use the stats tools for chart making.

Understanding how to work with both structured and unstructured data(think text in twitter). F500 Corporations work more with structured while tech work with both. There are endless algorithms for analyzing data. Start simple since you can nearly the same performance as the top of the line algorithms with simple ones. Linear regression, native bayes, k-means cluster, decision trees are more than enough. Maybe already up read up about bag of words with tfdf normalization if you work with text a lot. After that, most of the other algorithms will be based on those ones since it usual breaks down to supervised vs unsupervised learning with mixes in between.

Communicating data in a way that non-technical people can understand. This is not easy sometimes since you just spent 90%-95% of your time coding and analyzing. You might spend like 4%-9% on charts/presentation and like 1% of the time explaining what you did. But those last 5-10% are very important since that's how the client gets all their benefit out of your work. This includes even more technical data sciencists since they have to show the model works right.

A master's/phd in a STEM area also helps to establish credibility.

When you say one tool is apt for business and another for tech, what drives the difference?

A whore ain't nothing but a trick to a pimp. (Iceberg Slim)
Beauty is in the erection of the beholder. (duedue)
Grab your life by the pussy.
A better question to ask is "What EXACTLY do I want out of life and what EXACTLY am I doing to get EXACTLY that? If you can answer that question truthfully you will be the most Alpha motherfucker you will ever need to be. (PapayaTapper)
Reply
#32

Career in statistics/data science?

Quote: (06-24-2018 10:01 AM)duedue Wrote:  

Quote: (01-25-2015 03:33 PM)cibo Wrote:  

So I'm a data scientist and have worked for F500s and start ups.

Expanding on Agustin point. The data science field breaks down into:

Data analyst
Data Scientist
Data engineer

This creates a range between technical and business. Analysts have more business knowledge and have to do more client facing work while data engineers have to program more with less business facing. There's also a bit of difference in the tools between tech companies and normal corporations.

That said the common skills you need are:
SQL to get data from databases and is a fairly easy common language used for many of the Big data tools, Hiveql for hadoop, CQL for cassandra, etc. Most of the relational databases (mysql, teradata, orcale, sql server,etc.) are pretty much the same with some differences in functions and syntax here and there. The NoSQL databases have a lot of different types and can be a bit confusing if you don't understand relational databases enough. I'd work on those later after you have a solid grasp of Relational databases since the NoSQL's db's were created in response to the limitations of the traditional database.

Statistical programming language: SAS for corporate, R/python for technology/startups. This need somewhat depends the amount of raw statistical/machine learning work you're doing. Also when you start getting data from websites and apis, these start becoming more important.

Data visualization tool: Excel/VBA for coporate, d3.js for tech, and/or Tableau for any type of firm. You also use the stats tools for chart making.

Understanding how to work with both structured and unstructured data(think text in twitter). F500 Corporations work more with structured while tech work with both. There are endless algorithms for analyzing data. Start simple since you can nearly the same performance as the top of the line algorithms with simple ones. Linear regression, native bayes, k-means cluster, decision trees are more than enough. Maybe already up read up about bag of words with tfdf normalization if you work with text a lot. After that, most of the other algorithms will be based on those ones since it usual breaks down to supervised vs unsupervised learning with mixes in between.

Communicating data in a way that non-technical people can understand. This is not easy sometimes since you just spent 90%-95% of your time coding and analyzing. You might spend like 4%-9% on charts/presentation and like 1% of the time explaining what you did. But those last 5-10% are very important since that's how the client gets all their benefit out of your work. This includes even more technical data sciencists since they have to show the model works right.

A master's/phd in a STEM area also helps to establish credibility.

When you say one tool is apt for business and another for tech, what drives the difference?
Enterprise support, licence cost and level of technical sophistication.
Reply
#33

Career in statistics/data science?

Quote: (09-19-2014 07:50 PM)Peregrine Wrote:  

Quote: (09-19-2014 04:00 PM)polar Wrote:  

People in this field:

What are the top (3-5) specific skills, qualifications, software certificates, etc. that one can pick up to qualify for an entry-level role? Can someone drop a data sheet, particularly with recommendations on materials for self-study?

In order:

Math and/or statistics degree (BA, Masters, and/or PhD) from a reputable university
Excel/SQL/R/Matlab/etc and/or any programming language
Relevant experience in the industry that you want to analyze data in

Is a Masters/PhD in economics useful for getting a job as a data scientist? Assuming you have also learned the various programming languages used in the industry.
Reply
#34

Career in statistics/data science?

I’d say yes, but you should focus on courses in Econometrics and Statistics and things like Time Series and put them forward on your CV. Don’t only do Macroeconomics.
Reply
#35

Career in statistics/data science?

Quote: (04-06-2019 03:39 AM)Ouroboros Wrote:  

Quote: (09-19-2014 07:50 PM)Peregrine Wrote:  

Quote: (09-19-2014 04:00 PM)polar Wrote:  

People in this field:

What are the top (3-5) specific skills, qualifications, software certificates, etc. that one can pick up to qualify for an entry-level role? Can someone drop a data sheet, particularly with recommendations on materials for self-study?

In order:

Math and/or statistics degree (BA, Masters, and/or PhD) from a reputable university
Excel/SQL/R/Matlab/etc and/or any programming language
Relevant experience in the industry that you want to analyze data in

Is a Masters/PhD in economics useful for getting a job as a data scientist? Assuming you have also learned the various programming languages used in the industry.
You can get a job with phd in econ but its in what i consider to be more on the stats side. And thats fine since many people have full careers in that I area. But i would recommend you take an algorithm and data structure class along with a data base fundamentals class since most modern corporations have shittons of data in databases and engineers/it have no time to do data fetches for you.
Reply
#36

Career in statistics/data science?

The big problem with data science/analytics as a freelance/work from home/digital nomad style career is that most companies, definitely the bigger and better ones, will not outsource their personal data. It's important to them and makes up a foundation of their operations. They're not going to give that over to some guy on a beach in Thailand to work through.
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)