If you had a choice to learn SQL, R, or Python, for data analytics, which one would you pick to learn first and why? Want to learn something useful for financial analysis, policy analysis, ghg accounting, etc.
4
If you had a choice to learn SQL, R, or Python, for data analytics, which one would you pick to learn first and why? Want to learn something useful for financial analysis, policy analysis, ghg accounting, etc.
SQL first to extract data and do basic summarization filtering etc from databases. Then you layer either Python or R on top. Would not focus on both, just one as they serve similar functions with rich libraries.
I'd use python for the flexibility of working with full stack if necessary. I tend to use R to get something displayed quick but python for something that I put into production
I’ve been encouraged to learn Python! I have college SQL experience and it’s fairly easy to re-pick up with exercises. Friends in development have recommended Python for the versatility and popularity. So many people know it
Python - it has the most modern selection of tools and you can do far more than data analytics if you choose to do so. R is good if you work with old school analytics departments ie in the gov or old F500 companies. SQL you should know in conjunction with R and Python. It’s not good for analysis, it’s good for getting data out of databases. Once it’s out you use R or Python for analysis. SQL is also simple to learn compared to the other two.
Dm me if you’re good. :) we need really good engineers.
Unless you’re interested in heavy machine learning/advanced stats work, Python is more versatile than R. SQL is not really useful for data analysis.
Modern SQL is really powerful. Coming back to it after a few years I was shocked how much stuff I could move from python to sql. WITH function provides modularity. WINDOW is super powerful.
I prefer R. R is more statistically sound, ggplot2 is the best visualization package in any language, and shiny apps are good and easy to build. SQL is in addition to Python or R, not instead of. Python isn’t at all bad though.
I forgot about tidy in my response, really good for quick data wrangling. Better tool for consultants IMO for that reason. Also I prefer R Studio to any of the Python IDEs that I’ve used.
Our shop uses a lot of python though we aren’t pure DA
The real answer to this though is whichever is going to be easier for you to learn is the right one to get started in. If you’re going to go far in data science you’ll end up using both, even if you’re far more comfortable in one rather than the other. If you work in a group where some people use one or the other, get started in that one so you can ask questions and copy code. Stack overflow and other forums are great for asking questions but being able to ask a person you know questions at the start is way better plan. It is way harder to learn when you are just starting out, then after you’re a bit familiar you can read documentation and figure it out easy enough.
SQL u can learn in 2 weeks R in 2 weeks and python in 1 week. All programming languages have same shit some if else. Some GUI. If you have basics in OOP then it’s easy. Complex part of analytics is in more work around NLP complex quant modeling etc
Pwc2 thanks for sharing, I will try it out!
Just finished business school and many classes used R. After spending a few weeks really studying R (using ggplot2, tidy, etc), ifind it fairly easy to use and fun to learn.
What kind of sql? Is oracle sql still relevant? I know oracle real well but I’ve only worked in one industry so idk how popular it is outside my specific system
SQL + python/R
SQL should always be #1 on your list. After that it's down to preference - Python has a huge modern library, R is traditionally used in academic circles (Poli-Sci and Stats especially).
As a data scientist that knows them all here’s my two cents. 1. Python -> most versatile and “debuggable” 2. SQL -> for simple analytics tasks & reporting 3. R -> popular in academia, less and less in enterprise grade Disclaimer: I don’t know how deep and replicable your analyses need to be so even SQL with a good BI tool & bit of excel might be enough