Recently I’ve been working more with AI tools at work, primarly Cursor. It’s been extremely useful in automating the lower value parts of the analytics engineering workflow: writing config for data models (dbt yml files) and writing basic documentation in markdown (drawing from comments and logic in the SQL files). It’s also good enough to suggest adding basic tests, like standard unique and not_null dbt tests.

AI tools have made me more productive. They’ve also raised a question that keeps nagging at me: if I extrapolate this out 5-10 years, what’s left for humans to do? AI will move from the lower value to higher value tasks. Perhaps there’s some ceiling they’ll hit where humans will remain . But there’s concept in economics that offers a more nuanced—and perhaps surprising—perspective: comparative advantage.

Recently I’ve been auditing a Harvard course by Raj Chetty, “Using Big Data to Solve Economic and Social Problems”, freely available here.

It’s given me a lot to think about, both in the content of the lectures as well as how Chetty organizes the course and delivers the material.

Motivation

It’s the kind of economics class I think everyone should start with, and addresses the biggest complaint on the subject I have heard from new students to former econ majors like myself: too much theory and not enough impact on or connection to the real world. Chetty jumps straight into the impact and motivates learning the methods and theory of the course by offering them as tools to solve the biggest problems of our time: economic inequality and mobility, climate change, health, and so on.

After recently setting up a new personal laptop, I figured I’d collect a guide to the data science toolkit setup for the next time I need to set up a machine. But of course I didn’t actually do so.

Update 5/20/20: Well, that time has come earlier than expected, and I’m now setting up my old work computer as my personal computer (#uberon). I figured this time I might as well make a guide for myself while it’s fresh in my mind.

Nate Kornell’s research shows that difficulty in learning is often a sign that you’re learning. Difficulty is in fact desirable, and the right amount of it is key to effective, long-term learning. This is one of the most fascinating findings in my I learned in college but that I didn’t internalize until 7 years later, after working and learning many new skills.

Often when I’m learning a new skill or put into a new situation on the job, I encounter difficulty and a part of me thinks “this is a sign that I’m not good enough to do this.” The impostor syndrome is a well documented but not well internalized phenomenon that many of us feel, and certainly has something to do with that sentiment.

Having mostly taught myself R over the past few years, I wanted to gather some of the resources I’ve found most helpful in one place in order to help others who are interested in learning R as well.

R

A good place to get started with R and learn more is R studio cloud (https://rstudio.cloud/). You just make an account and access R Studio from the browser, no software to install and nothing to configure on your computer. And there’s lots of learning materials: interactive tutorials covering the basics of data science, cheatsheets for working with popular R packages and a guide to using RStudio Cloud.