What are good resources for learning about data science?
June 18, 2019 9:07 AM   Subscribe

I would like to learn more about data science to use it in my freelance marketing work. Do you have any recommendations for good resources I could use to do this?

I feel like I'm great at doing the "first look" analysis and reporting, but not so good at discovering deeper trends and phenomena. My primary goal would be to use all the data I get from my clients to come up with really interesting analysis. I'm interested in visualizing the data I get as well, so would like to learn about that (currently limited to PivotCharts in Excel). (My secondary goal is getting some highly-desired skills on my resume, so I'm inclined to lean toward more buzz-worthy technologies and techniques.)

My background:
Interested in books and online classes (e.g. Udemy, Udacity)
A lot of online ads experience, so I know the terminology
I've used Python in the past but am rather rusty. R as well. Have been dabbling in SQL.
Don't have a timeline to learn this stuff, but quicker is better
posted by miltthetank to Technology (9 answers total) 41 users marked this as a favorite
 
This came through my feed recently. I can't vouch for the New York Times, but I've heard good things about it.
posted by Glomar response at 9:39 AM on June 18, 2019


I posted about (and recommend) the online textbook R for Data Science in a recent thread. If you're more interested in Python, the (also free) Python Data Science Handbook is roughly equivalent. Both are great for working through at your own pace.

imo, R is a little friendlier (RStudio is amazing) and better at visualization. Python is more versatile, and probably the way to go if you're at all interested in getting into machine learning.
posted by theodolite at 10:19 AM on June 18, 2019 [5 favorites]


I strongly prefer working in R, especially for visualizations -- much nicer and more flexible than working even in seaborn in Python. I learned the basics as part of the Data Analysis nanodegree with Udacity, but honestly, the only way to get good is to use it. Either as part of work, or grab some open-source datasets and start exploring.

For more meta stuff, or introductions to techniques and stuff (among much else), I quite like Towards Data Science.
posted by kalimac at 11:26 AM on June 18, 2019


Kaggle: (Tutorials, free datasets and code you can download to experiment with and eventually competitions you can enter)

DataCamp: Some free lab-based courses (others you'll need to pay a membership fee for) to introduce you to R, Python, and SQL (among other languages) specifically geared at data science projects.
posted by Young Kullervo at 11:33 AM on June 18, 2019 [1 favorite]


Since you mention you're not great at discovering deeper trends quite yet, I do recommend brushing up on some basic statistics foundations for that. The book Practical Statistics for Data Scientists is really straightforward and not too jargony for that purpose. I've found that this foundation has really helped me recognize possibilities when I'm looking at data, which I can then explore further using Python or R.
posted by thebots at 12:09 PM on June 18, 2019 [7 favorites]


I found Python for Data Analysis to be invaluable, mostly because it's written by Wes McKinney, the guy who's the main dev for pandas. Jake VanderPlas is generally great as well, though, but I don't have experience with his handbook.

For resume-building, what you can do is most important, but all else equal, Python > R on that front.
posted by supercres at 1:11 PM on June 18, 2019 [1 favorite]


I second the advice to get some grounding in basic statistics first, if you don't already have some. (Data librarian person in higher ed here.)
posted by pantarei70 at 1:46 PM on June 18, 2019 [1 favorite]


Thanks for the great answers! I marked a few as "best" but all were very helpful. Next steps for me is getting Practical Stats for Data Scientists, cracking open a Python book, and signing up for Kaggle.
posted by miltthetank at 7:03 AM on June 19, 2019


Already linked in this question about learning data science, but although DataCamp is really awesome the company just recently attempted to cover up the CEO's very gross sexual harassment.

This really sucks, because it was a great way to affordably learn data science, but for me anyway it just feels really bad to support a company that endorses such sleazy/illegal behavior towards their employees.

DataCamp Sexual Harassment

DataCamp Teachers Boycott Their Own Classes Following Sexual Misconduct by Executive

A Multimillion-Dollar Startup Hid A Sexual Harassment Incident By Its CEO — Then A Community of Outsiders Dragged It Into the Light
posted by forkisbetter at 9:12 AM on June 19, 2019


« Older What to play on a Switch   |   Reasons to attend therapy more than once a week? Newer »

You are not logged in, either login or create an account to post comments