Help me crunch numbers...forever

December 4, 2013 6:26 AM

What are some potential career paths for someone who loves doing data analysis? What can I be doing right now to build my skills and experience?

For a while I've been kind of flailing around in terms of my career path. I have a bachelor's degree in linguistics and I work in entry-level nonprofit program management. I'm in a stable situation right now in terms of my job, but I always want to keep the future in mind so I can be aware of the kinds of professional development I want to be doing, and for the most part that future has been a huge blank.

Everyone keeps advising me to find my passion, and when every once in a while, I get the chance to work with data, numbers, surveys, polling, etc, I realize that turning a big mess of data (numerical or otherwise) into organized information that actually tells me something is totally the thing that I like doing. I love doing this at work when someone hands me a spreadsheet and asks me to make sense of it, I love doing it in my spare time with informal polling and graphing of information I'm interested in, and my reading list is full of data- and statistics-oriented blogs and articles.

I want to do this stuff every day. So: how do I get there from here? What types of careers involve a lot of number crunching? What background do people who do analysis typically have, and what can I do in my spare time to brush up on it? Looking further into the future, what kind of educational and professional goals should I have?

posted by capricorn to Work & Money (19 answers total)

Epidemiologist
posted by PuppetMcSockerson at 6:31 AM on December 4, 2013 [2 favorites]

I can see you working as a consultant if you also enjoy writing. After a couple years of that you could get an MBA and then continue in consulting or consider switching to finance. Just make sure you gravitate towards data heavy roles because there are other very different roles in these fields as well
posted by cacao at 6:35 AM on December 4, 2013

Designing surveys and analyzing survey data is its own thing - for educational goals down the line, you might consider courses or a graduate degree in survey methodology. That University of Maryland Joint Program in Survey Methodology I linked to also provides some information on opportunities and careers for people pursuing this path.

Biostatistics is another possibility - take a look at Johns Hopkins, and sign up for their Coursera courses to get a flavor of what that's like, if you haven't done so already.
posted by needled at 6:47 AM on December 4, 2013

I work with data analysts who do this type of thing. The ones I know have degrees in data analysis, decision science, and information systems. In addition to the title of data analyst, they have had job titles like statistical officer and information analyst. They've worked in a variety of industries: health care, payments, basically anywhere that companies collect data and want to use it to answer business questions and identify trends.

Have you ever used Tableau? There is a free version for personal use. It's a data visualization tool, but I think it's also an interesting way to manipulate with big data sets in a way that you can't really do with a program like Excel. There are free public data sets available online, which is nice for playing around.
posted by neushoorn at 6:52 AM on December 4, 2013 [4 favorites]

If you enjoy working in a college/university setting, you might want to explore institutional research -- the Association for Institutional Research website will give you a lot more info. Higher ed is becoming more and more focused on assessment and outcomes, and people who can do this kind of work will, I think, be finding more employment possibilities.
posted by Kat Allison at 7:03 AM on December 4, 2013 [2 favorites]

I know people in web analytics/User Experience/usability/Information Architecture/web marketing analysis who do quite a bit of that. I work on a particular accounting/ERP software package and many of my customers have at least one person who does data crunching, informally or formally (if it's formally they usually fall under the Business Analyst title, though that means a lot of things), and some of my biggest customers have actual financial reporting departments.

There's also Human Factors, which is what I would drop everything to do if someone handed me a Go To Grad School Free card.
posted by Lyn Never at 7:04 AM on December 4, 2013 [2 favorites]

Best answer: Hi, I am living the dream as a data analyst. My art history degree certainly didn't do much to prepare me but I have taught myself a number of skills. The foundational skill that has enabled me to use big corporate systems you can't easily train yourself on (like SAS or Business Objects) is SQL. Once you know how relational databases work you can figure out most reporting software.

Another fun tool that is being used more is R: a programming language for data analysis. You can check it out at sites like Code Academy.

Feel free to memail me.
posted by munchingzombie at 7:13 AM on December 4, 2013 [14 favorites]

I'm a Sales Operations Analyst. I'm self taught, but I do have an MBA.

I'm an Excel Wizard. That's pretty straightforward to learn, lots of tutorials on-line, and communities to hang out in.

My specialty is CRM (customer relationship management) for sales. I use a niche system now, but Salesforce.com is the one system I suggest getting your hands dirty with. It's ubiquitous and intuitive and easy to learn and use.

You don't have to major in anything, so much of it is Excel and an understanding of the story that data tells.
posted by Ruthless Bunny at 7:26 AM on December 4, 2013 [4 favorites]

Actuarial science. That's where the rubber meets the road in terms of using statistics in immediately and concretely practical ways. An actuarial trade group has a page on what it means to be an actuary and how to become one.

Actuaries aren't commonly heard of, but their influence is enormous. Say you're trying to do something about your insurance policy with your agent.* "Gee," the agent says, "I'll have to ask Underwriting about that." Well Underwriting lives in as much fear of Actuarial as your agent does of Underwriting. Underwriting basically just implements what Actuarial comes up with, and it's all number crunching.

It's also a career that pays well and is highly in demand, partly because you need not only a college degree but a passing grade on a series of wickedly difficult professional examinations. But if crunching numbers is your thing, well, this probably is too.

*It's not only insurance, just mostly. Non-insurance financial institutions, including pension plans and public benefits programs, rely heavily on their actuaries.
posted by valkyryn at 7:27 AM on December 4, 2013 [4 favorites]

This is basically my job (sort of). I am a consulting data analyst. I do a lot of work with Excel/SQL and also Tableau (which is excellent)

You certainly should learn some SQL (try GalaxQL as a tutorial) because it is a corner stone of dealing with large datasets. However (as my askme profile shows) you can probably wing it for a bit with minimal training.

If you are really feeling adventurous look into R ,which none of us here know, and we all sort of wish we did.

Playing around with scripting languages will be useful because getting your data into a shape where you can analyse it is usually a fairly big bit of the work, and there is nothing like a good scripting language to mess around with a few folders of files and get them ready for inserting into SQL or Excel. Try Python and Powershell.

For the same reasons a fair understanding of XML would not go amiss.

Perhaps you could find things in your current job that use these skills. You can then start tilting your CV to these sorts of areas and moving into more and more data heavy jobs.
posted by Just this guy, y'know at 7:55 AM on December 4, 2013 [4 favorites]

I'm in an Educational Psychology - Statistics and Measurement graduate program, for about the same reasons you articulate in your question. My background before this was a doctoral program in pure math (but despite how much I wanted it to, pure math research just didn't do it for me), then a master's program in education (in which I learned a lot, but couldn't deal with how much everyone else in the program hated anything to do with math or numbers), and I finally settled on where I am now, which is a good mix of getting to be part of the education world while also getting to do the actual type of work I enjoy.

You might try checking out Educational Psychology MS programs in your area to see if any of them have Statistics and Measurement concentrations. Even though my program is technically in education, our students go on to do statistics and data analysis work in a variety of fields, and unlike a lot of other academic areas, it's got a pretty good job market at the moment because it's a pretty niche field (so not a ton of students in it nationally) with applications in a wide variety of public and private industry jobs. The opportunities for graduate student funding in my program are also much better than the opportunities in other School of Ed programs at my university, so don't let the overall funding climate for higher ed deter you from checking into this type of graduate program. Good luck! If you'd like to know more about my program in particular or get in touch with some professors in the field to learn more, feel free to memail me.
posted by augustimagination at 9:32 AM on December 4, 2013 [2 favorites]

Best answer: Definitely learn SQL. Whatever you do, don't use MS Access as your learning environment. Download MySQL and Toad for MySQL and start working through any of the number of excellent T-SQL books out there. If you like sports (particularly baseball) there is a ton of data available if you know how to use web scraping scripts with R offering one of the greatest MLB related libraries ever - pitchRx.

When I was looking for a new job a few years ago I used the book Baseball Hacks to sharpen up my stat skills that had went to pasture thanks to years of being in management. Some of the examples (the environment used is R and MySQL) are no longer relevant thanks to things like pitchRx, but there are a lot of good examples to work through. The key to learning R and MySQL is to find a project you'd like to do and then start figuring out if there's a data source available. You can always consult Data.gov for government data sets as well.

Here's another book I'm currently working through to prep for next fantasy baseball season - Mathletics.
posted by playertobenamedlater at 10:23 AM on December 4, 2013 [7 favorites]

Best answer: I'm also living the dream. Learn SQL. If you are already in the nonprofit sector, you may find interest in data analysis of fundraising databases, to help nonprofits raise money more efficiently. Those jobs are out there, the keyword you're looking for is "Advancement Services." Check out the AASP. The analysts are always in very high demand, and demand is increasing as nonprofits continue to figure out how to do more with less.
posted by juniperesque at 10:43 AM on December 4, 2013

Best answer: More on learning SQL - here is a good interactive site for practice:

http://sqlzoo.net/wiki/Main_Page
posted by gregjunior at 10:45 AM on December 4, 2013

SQL, R, and maybe Python are the big tools of this trade. If you like sports a lot of people are doing cool statistically analyses there these days.
posted by Aizkolari at 1:47 PM on December 4, 2013

Best answer: I'm a Senior Data Scientist at a social gaming company. I play with data all day every day, splitting my time between architecture building and analysis 50/50 or so.

SQL is critical; I've been a certified DBA in Oracle and Informix and built plenty of data driven applications in the past and thought I knew SQL pretty well... until I started this job and realized I know nothing. There's not really an upper bound to how well you can know SQL for a job like this. I get questions all the time like, "How much money do devices spend that have logged in between 25 and 35 days since we first saw them that have also played X game as their first game and Y game at least a dozen times in their first week and have played Z game since we released it this week?" An analyst will take a ticket and have it by the end of the week, a senior person will do it with you hovering over his/her shoulder.

Aside from that I use R almost daily for visualization and analysis. It's great to be able to run a query and throw the data onto a world map or whatever with a single command. It's replaced excel as my goto app for prettifying data. The learning curve is steep, but a couple days of intensive study will get you reasonably proficient for at least basic visualization. If you know enough statistics, the analysis tools R has are great. I wish I knew more statistics... advanced stats in my undergrad only barely scratched the surface.

Most architecture stuff I do in pure SQL, kicked off by cron or an ETL script - which are all written in Python. I've written some of it, but the guys that are responsible for the data importing do a lot more of that.

To summarize all this, I would say: learn the hell out of SQL, and get as firm a grasp on statistics as you can, and you should be primed to at least get a foot in the door.
posted by skintension at 3:05 PM on December 4, 2013 [4 favorites]

Best answer: I do this in advertising (campaign effectiveness reporting). Skills involved are:
SQL and relational databases - 90% of my day usually
R
Python with Pandas / Numpy
Off-the-Shelf report builders (Microsoft Reporting Services, Crystal Reports, etc)

Rudimentary statistics (seriously most of it is just counting and simple ratios, maybe a standard deviation or correlation coefficient at the most)

Major bonus points for :
Any sort of javascript canvas/svg library (I like d3)

Machine Learning:
Classification
Clustering
Feature identification
posted by spatula at 4:56 PM on December 4, 2013 [1 favorite]

Best answer: Glad someone mentioned JS and D3. Part of my job now has me dealing with JS and D3 daily and, once I got over my disdain for JS, have really grown to love it. But, and I say this not to dissuade you in any way shape or form, I'd put learning those skills down at the bottom of the priority list and put SQL at the very, very top since you really need to have a good understanding of relational databases (both how they're structured and implemented and how they store data) and be a SME (subject matter expert on the data) before any of those tools are really useful or applicable.

Once you have a good understanding of SQL and want to jump into R (something I would recommend over Python) you want to start working with Norm Matloff's The Art of R Programming. There's a draft of it on his UC Davis site along with a ton of other documents you might find interesting. Here's the ECS 132, Data Modeling and Analysis course site that he taught this fall at UC Davis. Might be a bit too advanced for a starting point, but once you're comfortable with R this is another great resource from someone that really gets R and how to get other to use it.
posted by playertobenamedlater at 8:13 AM on December 5, 2013 [2 favorites]

Best answer: You've described my job. I'm a survey research analyst with a Ph.D. in social psychology. Echoing the calls to learn SQL and R--I don't use them myself but I do see job listings that want people with that experience. At the very least, learn how to use SAS and SPSS and take as many stats courses as possible if you decide to go to grad school. It's a fun job with a lot of variety in terms of the topics you survey people on. Memail me if you are interested in learning more.
posted by Fuego at 1:56 PM on December 5, 2013 [1 favorite]