Trying to Switch From Qualitative Field into Statistics

#1
Hello everyone, I've always been into statistics but I never studied it, other than a little in high school.
I've decided I want to try to get a job in the 'big data' field, or any job where I would analyze data in ways that enable business decisions, for example being the guy at Facebook who decides where to spend ad money etc.
Clearly, for this I need to get myself a background in statistics and I'm willing to devote a few years to this. However, as I stated above, I really don't have a background in statistics or anything quantitative. I went to college in Canada (McGill University) where I majored in Political Science. I graduated in 2008. My GPA wasn't stellar, probably around a 3.0 or a little lower.
So I want to know what my options are in terms of acquiring the applicable knowledge and eventually turning that into a career.
It'll be an uphill battle but I'm willing to work hard.
 

dss

New Member
#2
Hi,

I think you are in a similar situation as I am. I have started with a job which requires a lot of statistical analysis but my background is in computer science engineering. After having looked at many different forums and books, I finally came to this book which I found very useful for a beginner. It’s a book called ‘Statistics’ by Freedman, Pisani, & Purves. My belief is that it can’t be any easier that this book. They talk about statistics in concepts, rather than equations. Have a look at its table of contents.

I hope it helps you. And in the mean time, if you yourself find any other books/helpful advice, do let me know.

Cheers,
DSS
 

spunky

Doesn't actually exist
#3
I went to college in Canada (McGill University) where I majored in Political Science
tell us a little bit more about your background in computer programming? because working with Big Data (and actually, doing any kind of serious statistical analysis) will require you to become pretty dexterous at computer programming (ideally, in R, in my biased opinion).

if i were you and had no background in statistics and no background in computer science (plus knowing what i know today) i would seriously consider going back to school and getting at least an associate's degree in something computer science/math/stats related, at least to get you started.
 

noetsi

Fortran must die
#4
I recently earned a master's degree in measurement and statistics from a state university, in part because I do data analysis, but learned much of what I know outside universities. My conclusion from that is that, unless you have lots of free time and probably even then, there are limits to what you can learn in classes. Academic classes don't focus as much as they might on details about diagnostics, going from an article to actually generating that information in your specific software, and acquiring and formating/transforming your data. Also you will only learn a few methods this way, many of which you will not use regularly (while ones you want to learn won't be offered...).

Reading text, on line articles, and comming to places like this is probably more useful. Of course my program was not in the stats department so maybe it is different there (if you have the math background to attend such corses). I spend much of my day downloading class lectures and on line articles and then comming here and asking the experts "is this right" or asking about stuff left out of the course.

From my own research experience qualitative methods are entirely different than quantitative ones. They really need to be learned by doing. You will be lucky to find many classes on them.

I agree with spunky about having to learn to use computer software (and probably SQL and on line searches which takes quite a while to learn). But for most of what you will do (unless it is at an extremely high level) it won't much matter if you use SAS, R, SPSS, or STRATA. They aren't that different until you reach the stratosphere.....
 

spunky

Doesn't actually exist
#5
But for most of what you will do (unless it is at an extremely high level) it won't much matter if you use SAS, R, SPSS, or STRATA. They aren't that different until you reach the stratosphere.....
well... the OP does seem to aim for the 'stratosphere' as you phrase it. just look at this paragraph from his/her thread:

"job in the 'big data' field, or any job where I would analyze data in ways that enable business decisions, for example being the guy at Facebook who decides where to spend ad money "

everytime 'BigData' pops up you immediatley know you're going to need a repertoire of highly-specialised technical skills. sometimes it's not even so much statistics as it is computer programming and database managing. that's why i inquired about the OP's programming experience because, if it is none, then there's a long, long (but wonderful) road ahead waiting for him/her.
 

noetsi

Fortran must die
#6
I did (do) those kind of jobs and worked with people who did it at an even higher level in reasonable sized companies and government. I disagree that it mattered which software they used (in point of fact most used excel not statistical software to make financial analysis).

This is an ongoing disagreement between us spunky, I don't think the way financial analysis is being done in most companies is the way you think it is being done. Until they come up with hard numbers what software financial analysts use (which is what I think the OP means) it will likely stay the same.
 

spunky

Doesn't actually exist
#7
This is an ongoing disagreement between us spunky, I don't think the way financial analysis is being done in most companies is the way you think it is being done. Until they come up with hard numbers what software financial analysts use (which is what I think the OP means) it will likely stay the same.
the inquiry is not about financial analysis, it's about training in BigData and analytics. because i totally agree with you: nobody needs to be a computer expert just to become a generic financial analyst for a company. HOWEVER, if you want to dabble in the world of BigData then *yes*, you do need technical training beyond what Excel lets you do.

i think we're now experiencing what happened when that TV show CSI became super popular and suddenly EVERYONE wanted to become a forensics analyst. for some reason, now and again it is being touted that "BigData is the future" and "training in BigData is the way to innovation", BigData this, BigData that... so i'm not surprised that suddenly more and more people are flocking towards gaining technical skills in order to secure some form of employment.

i mean, in just one week this is already our 2nd inquiry on this stuff :D
 

noetsi

Fortran must die
#8
I suppose it depends on what the OP means by bigdata. My sense is that in most organizations that means financial analysis not the type of statistics we discuss here commonly.
 

Mean Joe

TS Contributor
#10
OP is from 3 months ago.

You don't need to go to school first to get a job. It might be possible to get an "entry level" position at the company first, with the understanding that you'd be learning at the same time to get into the desired position in the company.
 
#12
To get started, I would suggest learning how to program first. I would recommend python as the first language to learn to get started. It's popular in the scientific world and it reads a lot like pseudo code or what you might say if you described a problem step by step in english. Try the learnpythonthehardway tutorial.

Once you are comfortable with python, I would suggest moving on to ThinkStats or some of Allen Downey's other books. ThinkStats is a great introduction to statistics. It uses programming in python to help the student develop an intuition about the subject using a medium (programming) that they are familiar with. It will help you become a better programmer as you learn more about statistics. A win win.

Good luck.
 
#13
you said that you have a background of Political Science and you want to go to the field of statistic risk checking and analyzing company data to indicate the correct path of further movement

1.one idea can be Six sigma training ASQ is one of the institutes
2.ofcourse many people have given you idea on having online classes on states
3.take some near by home coaching tutorial for beginner stats so that you get hold on basic idea & concepts & terminologies & confidence & span of the subject
4.then joine some online course.
5.try and attempt easy case studies
6.join a 3 month or six months program on statistics form the institute like ASQ and other also you will find as you move on your way

all the best and good luck
 
#14
I am pretty much in a similar boat. I was the one who asked the question about Big Data Analytics and I am trying to figure out how I can go from the tax accounting industry to move into Big Data/Data Science, whatever you want to call it.

Are any of you guys familiar with this online program for Data Analytics, any thoughts on it?

http://www.statistics.com/data-analytics1#program

Any thoughts as to the courses in the program. Does it seem to you guys that you can get a good statistics background going through this and chances for employment will increase or not?
 

spunky

Doesn't actually exist
#15
Any thoughts as to the courses in the program. Does it seem to you guys that you can get a good statistics background going through this and chances for employment will increase or not?
there's a really smart member of this forum bryangoodrich who works in Big Data and Big Data-like problems. maybe you could send him a PM asking for his opinion? he got me really interested in it and now i'm seriously considering to actually dumping academia and pursuing a job in Big Data as well once i finish my PhD
 

noetsi

Fortran must die
#16
Given how few professorships there are you are likely to work in bigdata (or as I refer to it, the "real world') rather than academics a good part of your career spunky. Of course I still don't really know what that term means :p
 

spunky

Doesn't actually exist
#19
I assumed he meant a specific type of job and/or analysis not a specific type of data.
that's exactly the kind of job he was inquiring about. jobs related to the analysis of Big Data.

among quantitative folks, anyone can just analyse "data", it's not terribly interesting to ask about it because anyone, in that group, can do it. but not everyone can analyse Big Data. the challenges implicit in it are what separate it from everyday data analysis folks.

you can also ask bryangoodrich. his current job is about that, working with Big Data and the issues involved with it. he's the one who got me into this stuff :)
 

noetsi

Fortran must die
#20
I imagine the problem with big data was discovered by the same people who "solved" the Y2k problem (sic).:p

My cynism, or perhaps ignorance, showing again.