Help me data better

October 4, 2021 2:55 PM

What should I learn if I'm interested into breaking into data analyst roles in the nonprofit space?

I have been a program manager for years in nonprofits, primarily in education, community development and healthcare. I have always enjoyed the data part of my roles and I'd like to explore a move more in that direction while staying in the nonprofit field, but I think I need to beef up some of my technical skills. Here's a few roles I've come across recently to serve as illustrative examples of jobs I'd potentially be interested in: 1, 2

Currently I am quite comfortable with most Excel stuff and decently comfortable with Tableau and Google data studio. I pick up using basic database functions and building reports in databases pretty quickly (airtable, salesforce) but I'm not an expert at anything.

I have no programming experience (SQL, python, R etc-- basically starting from zero.)

What should I focus on learning? I'd like to prioritize things that are: versatile, eas(ier) to learn, and most likely to be used by nonprofits.

Here's where I'm leaning based on my research so far, would appreciate thoughts:

-R: It seems like R is often requested and probably easier to learn than python? Or should I just start with STATA and see how that goes?
-Maybe SQL? Feels like it would be useful, in a different way than R, but I'm not sure if it's really necessary for the kinds of roles I'm looking at.
-Tableau: Since I have a start with tableau, it seems like that would be a good bet for the visualization side of things.
-I should probably take a basic stats class of some sort to make sure I'm solid on the math concepts.

What do you all think? I'm also all ears for any other advice on how to approach this (learning tips, etc).

Thanks!

posted by geegollygosh to Technology (17 answers total)

SQL
posted by sandmanwv at 3:10 PM on October 4, 2021 [2 favorites]

Best answer: From my perspective as a former analyst and now data engineer-type, I view SQL as by *far* the most critical skill for anyone in the data space. In most real world scenarios, the data you’ll be analyzing will live in a database, and SQL will be the key to finding what you need and putting it together into something useful. Then, maybe you need a Python or R to explore it further, but in many cases you’ll be able to extract insights with SQL and Tableau or Excel alone. If I’m hiring someone, SQL skills are necessary; Python is secondary, and R is… eh. R or Python only, for me, screams “academic/recent data science certificate grad who probably can’t do much of practical value.”
posted by twigatwig at 3:17 PM on October 4, 2021 [4 favorites]

Best answer: Hello! I'm a data scientist, and I have several staff from non data backgrounds that I'm aiming to develop, so I can pass on to you the resources that I'm providing them.

SQL
Here's a fun resource to get you started with SQL.
SQL is going to be very useful because most data is going to be in a SQL database, so knowing your way around at the least will be helpful.
I would for sure learn it.

R
I love R. R is my favourite language.
I think Python is probably more versatile and popular, but the data science community within R is extremely welcoming and helpful and I find the learning resources are better.
I sent our new graduate this and asked her to go through chapter 1 to get doing with R and R Studio,

And also this set of primer classes to get her used to core concepts.
There are more resources here

I have many many many more resources and I'll happily share them as and when you like. Feel free to memail if there's specifics that you're interested in.
I would also look at tidy tuesday. This is a weekly data science challenge, whcih is a great way to play around with new data sets and get practice in.


On preview:
R vs Python vs SQL.
Learn SQL first. You unquestionably need to know it at least a bit.
R vis Python will depend on your industry. I find that medical stuff and more scientific focussed areas tend to R and businessy stuff tends more to python. I think Python with Pandas (the usual data science library) is basically a poor man's version of R (but this is very contentious and most people will probably, and rightly, disagree with me)
posted by Just this guy, y'know at 3:26 PM on October 4, 2021 [9 favorites]

Best answer: I wouldn't bother with STATA. While it's "easier" in many respects than Python or R, it's not free, and the programming lessons learned are not always directly applicable to Python or R.

Agree that Python/ R is industry specific. One generalization about the difference between the two is that R started off as a language for statistics (by statisticians). Other "programming aspects" were sort of kludged in later. On the other hand, Python started off as a general purpose programming language for programmers. The statistics part was sort of kludged in later.

If it helps at all, an analogous "controversy" might be the mac vs. pc debate ~15 years ago. There are strong opinions on both sides.

Besides looking at job posts (to see what organizations in your industry are looking for), you might also want to think about what specific types of data/ data sets you would be most interested in using. For instance, if you want to work with NHANES data, you may find that R is better supported than python.

If you are open to a paid approach, you could do worse than using Datacamp.com.
posted by oceano at 3:57 PM on October 4, 2021 [2 favorites]

Oh and I didn't learn it in that order, but I do agree that SQL might be a good "first programming language." The syntax is pretty straightforward for a programming language. In addition, I think it would force you to get comfortable with manipulating data in a relatively structured way.
posted by oceano at 4:01 PM on October 4, 2021

Could I politely dissuade you away from datacamp.
They have a history of being extremely sketchy.
Some details here.

A lot of people who delivered training material for them have publicly asked folk not to use the courses that they had written.
posted by Just this guy, y'know at 4:03 PM on October 4, 2021 [2 favorites]

I did a six month Data Analysis certificate aimed at working professionals (two nights a week). It covered basically all of the technologies you mention and it opened a startlingly large number of doors for me.

I highly recommend something similar.
posted by Tell Me No Lies at 4:34 PM on October 4, 2021 [3 favorites]

Best answer: Absolutely learn SQL! And r or python! In the non profit world, few places will choose to afford Stata. Tableau might be more of a thing in some places too.

Also, those two jobs you linked are very different. Evaluation is a field in itself, and no amount of data skills is going to get you into that unless you do some work in qualitative analysis.
posted by advicepig at 4:39 PM on October 4, 2021 [2 favorites]

Best answer: The beginning stages of learning programming can be very challenging. It’s okay if things don’t make sense at first! Persistence can really pay off! Consider it more of a marathon rather than a sprint.

Personally I was drawn to Python as my first language. I found the syntax easier to understand and there is a lot of fun and engaging material available online in addition to a plethora of books. Python is seen by many as a great beginner language and many colleges have already changed their Computer Science curriculum to teach it as the first language.

I would take a look at some introductory Python, R, and SQL material and see what makes the most sense to you. Then go with that. Either way you will probably want to learn at least some SQL.
posted by mundo at 9:57 PM on October 4, 2021

Best answer: Hello, I am a data analyst at a non-profit! The things I use the most are SQL, PowerBI (Tableau to a lesser extent, since the pricing is higher, but skills from one transfer well to the other), and Excel. Many non-profits will not have the funds for more specific or advanced tools.

It may be worth getting some experience in any master data management (MDM) or data warehousing tool, because non-profits tend to have a bunch of different applications and data stores that you need to mash together. Again, the skills cross over.

I need to learn R - if I were hiring right now, it wouldn’t be a requirement, but it would be a point in your favor.

If you can only do one thing, or if you want to start with what will make you a more attractive candidate ASAP, though, really, you want to learn SQL.

Message me if you want to talk more about this. I love my job and think that non-profits desperately need more data people. Also, I might be able to send you some leads.
posted by punchtothehead at 3:39 AM on October 5, 2021 [3 favorites]

Oh and something to keep in mind is that intro stat classes sometimes use a particular software (e.g. R, SPSS, Excel). In addition, even within the same school, the stats 101 class may use one software, but the stats for x major (e.g. psych) may use a different one.
posted by oceano at 6:06 AM on October 5, 2021

Best answer: Outside suggestions: Jupyter Notebooks and Julia.
Jupyter Notebooks provide a place for notebooks where you explore and understand the data and its structure. SQL has tools like Oracle SQL Developer and MS SQL Server Manager for scripting up repeated actions on your data. Julia is a young language that uses optimal libraries (Fortran has historic efficacy) and clear expression of the maths manipulating your data.

Think about the process of building a mental model of the data from its types (numbers, names, dates) and how they interact. Grow your skills in that, and get experience with tools to amplify those skills.
posted by k3ninho at 8:20 AM on October 5, 2021

Response by poster: Thank you everyone--super helpful answers! I'll get started with SQL and go from there.

Will also probably memail a few of you who offered :)
posted by geegollygosh at 10:28 AM on October 5, 2021 [1 favorite]

I am in a similar spot to you, leaning into my love for spreadsheets and experience working in international development. Thanks to multiple folks here on AskMe, I started poking around GalaXQL last week, and found it well-structured and enjoyable to work through.

Tableau is popular in job ads, and whatever skills you have/develop there would probably translate easily into other platforms, if an employer you were applying to uses a different platform.

R is used by a lot of folks for data cleaning/munging and visualization as well as analysis, and so it seems like a good bet for a more multi-purpose tool to learn than STATA. It is also approved by the federal government (whereas, e.g., STATA is not), so it's a good one to have if you're ever looking at fed jobs.

As an M&E professional, I wouldn't recommend applying for an evaluator job coming from program management without being able to demonstrate a solid background in the theory and practice of evaluation. However, I think building your understanding of evaluation would be incredibly helpful to do better work with data, period. The American Evaluation Association has a 101 course that might be of interest.

I'd love to chat with you and punchtothehead, if you'd be interested!
posted by rrrrrrrrrt at 1:56 PM on October 5, 2021 [1 favorite]

For a basic SQL tutorial, take a look at galaxql. It's a little dated, but it's free and will give you the basics for commands. Some of it can be skipped: from a data analysis point of view, you don't need the lessons in dropping and altering tables, at least not initially. It's cute, has a neat visual and is a good place to learn a the basics.
posted by Hactar at 10:07 PM on October 5, 2021

Following this! I work in evaluation in the non-profit field, which seems slightly different than the data side based on some answers to this post, but we definitely use R too! I'd love to chat with you or some of the others in the comments of this post - feel free to memail me any of you!

If you’re interested in learning a bit more about evaluation too I’d recommend looking into learning survey-writing skills.
posted by azalea_chant at 9:45 AM on October 6, 2021 [1 favorite]

Would it be sensible or useful for me to set up a Discord server for those involved or just interested in Data Science.

Hmm, that's basically quite easy to do so I will do it anyway.

HERE is an invite link for a MeFi Datascience chat for if you want to learn data science type things
posted by Just this guy, y'know at 2:02 AM on October 7, 2021 [4 favorites]