Data/Research TA Resources

General

• Cheatsheet of R commands for data work .
  • Data cleaning: dplyr and tidyr
  • Figures: ggplot2
  • Working with character/word vectors: stringr
  • If you plan on using R Notebooks: rmarkdown
  • Original source
• Cheatsheet of Stata commands for data work
  • Original source

Workshop 1: script-writing

• See my econometrics notes below for a refresher on R, R Notebooks, and for an intro econometrics refresher.
• R: fixest for running regressions (as opposed to lm or plm)
  • faster estimation
  • flexible formula writing allows simultaneous estimation of many similar regression models, much simpler and more convenient inclusion of leads, lags, and differences of variables, and intuitive specification of standard errors
  • its etable function provides a nice pipeline for creating very customizable tables automatically (better in my opinion than the usual suggestion, which is the stargazer package)
• R: magrittr for very intuitive and readable script-writing especially for data processing. An intro here .
• Stata: an intro to data cleaning functions

Workshop 2: project-oriented workflows

• A guide to code and data for researchers

R-specific:

• Why use R Projects? (the other chapters here are great too)
  • A more basic guide here
• Relative filepaths using the here package
• A guide to R Notebooks
• A bit more advanced: R profiles
• Setting seeds for replicability

Stata-specific:

• Stata’s equivalent of R’s here package for relative filepaths
• Workflows for automating tables
• More advanced: Stata’s equivalent of R Profiles
• More advanced: Jupyter Notebooks for Stata
• Setting seeds for replicability

Additional topics

Topics I’d cover with more time

• Debugging tips: how to identify bugs in your code
• Using ChatGPT as a coding resource. It is quite error prone (especially for Stata since it isn’t open source) but invaluable when it does work. Some things I use it for:
  • “How do I do [task] in Stata?”
  • “How do I implement [task] in R using [package, e.g. the tidyverse]?”
  • Copy and paste a chunk of code and then ask what each line is doing
  • [pasted code] What’s a more efficient way of accomplishing the same thing?"
  • “This is my code: [pasted code]. I get an error that says [error]. Where is my mistake?”
• Writing your own functions
  • A guide in R
  • A guide in Python
• Implementing different kinds of regressions (I do this a little bit in my econometrics notes)
• Customizing regression tables
• Data visualization
  • R: see the ggplot2 cheatsheet
  • Stata: see the commands in the cheatsheet and here ’s a guided introduction

Other possibly helpful research resources:

• Trello for keeping notes, maintaining to-do lists, storing relevant documents, summarizing research meetings, etc. Sychronizes across all devices.
• How to present an applied micro paper
• Browser extensions
  • EZProxy Redirect to access online resources that Columbia has subscriptions to when away from university internet and without using a VPN
  • Simple Mass Downloader to download all files contained in a web page
• Literature review
  • Google Scholar search is obvious but also click on the “Cited by” link under a search result to find other relevant and possibly more up to date papers and methods or see if someone’s already done what you want to do