Keeping track of your code with Git and GitHub

EVR 628- Intro to Environmental Data Science

Juan Carlos Villaseñor-Derbez (JC)

Rosenstiel School of Marine, Atmospheric & Earth Science and Institute for Data Science & Computing

Assigned Readings

Lowndes et al. (2017)
Braga et al. (2023)
Think of the last figure, table, or number you produced for a class or job. If I asked you to regenerate it today with one column of the raw data changed, could you? What exactly would break?
Why isn’t DropBox or GoogleDrive enough?
Who is reproducibility for?

Learning Objectives

By the end of this week, you should be able to:

Understand the importance of reproducibility
Create the basic structure for an organized project
Explain the difference between Git and GitHub
Create and “clone” a “repo”
“commit” and “push”

Learning Git and GitHub may (will) be painful

Usually taught later in courses
Limited use if you learn it at the end
I would rather push materials to later in the course, but have you dominate this

Today’s Agenda

Reproducible Research
File Structure & Organization
Version Control with git
GitHub & Collaboration
Hands-on: Building Your Repository

Version control and reproducibility

Why Should You Care?

Career advancement: Version control is a required skill in data science
- Example job postings industry, academia, and government
Collaboration: Work with others without breaking their code
Reproducibility: Essential for scientific integrity / managing the real world
Maintain versions through time
Backup & recovery: Never lose your work again
Portfolio building: Showcase your skills to employers (tuna data by E. Rodriguez)

The Problem: Research can be chaotic

File naming: analysis_final_v2_really_final_USE_THIS_ONE.R
- (And the filenames loaded therein might look even worse)
Lost work: “Where did I put that working version?”
Collaboration conflicts: “Who changed what and when?”
Reproducibility failures: “Why do we get different results?”
Publication delays: “I can’t reproduce my own results”

Reproducible Research

Research that can be repeated by others using the same data and methods

Key Components:

Data: Raw, processed, and metadata
Code: Scripts, functions, and workflows
Documentation: Methods, assumptions, and decisions
Environment: Software versions and dependencies
Results: Outputs, tables, figures

The Reproducibility Crisis

70% of researchers have tried and failed to reproduce another scientist’s experiments (Baker 2016)
50% of researchers have failed to reproduce their own experiments (Baker 2016)
$28 billion wasted annually on irreproducible pre-clinical research (Freedman et al. 2016)
Environmental science is not immune to these issues

Real-World Example: Climate Data Analysis

Scenario: You’re analyzing sea surface temperature trends

Without version control:

Multiple versions of scripts with unclear differences
No record of data processing steps
Can’t track which parameters produced which results
Collaboration becomes chaotic

With version control:

Clear history of all changes
Reproducible workflow from raw data to final results
Easy collaboration and peer review
Confidence in your results

My drawer folder of shame

Who did what?
When?
Why?

Organizing your projects

File vs path

Folder $\neq$ file

Folder is an address (where in your computer does this file live?)
File name is the actual file
When reading / writing data from R, we must specify where the file lives

Folders are represented as my_folder/
Files always have an .extension, which tells us the type of file

Project Structure Fundamentals

Important

Good file organization is the foundation of reproducible research

Principles:

Clarity: Names should be self-explanatory
Separation: Keep raw data, processed data, and outputs separate
Consistency: Respect folder structure
Documentation: Include README files and metadata

Recommended Project Structure

cool_project/
├── data/
│   ├── raw/              # Original, unmodified data
│   ├── processed/        # Cleaned, transformed data
|   └── output/           # Output from your analyses
├── scripts/
│   ├── 01_processing/
│   ├── 02_analysis/
│   └── 03_contents/
├── results/
│   ├── figures/
│   └── tables/
├── docs/                 # Documents (optional)
├── slides/               # Slides (optional)
├── coool_project.Rproj   # Rstudio project (MANDATORY)
├── README.md             # Project overview (MANDATORY)
└── .gitignore            # Files to exclude from version control

Example

What’s missing?

File Naming Best Practices

DO:

Use descriptive names: sea_surface_temp_analysis.R
- I prefer snake_case, but CamelCase is appropriate too
Use consistent separators: underscores or hyphens

DON’T:

Use spaces: my file.R
Use special characters: analysis@final.R
Use vague names: stuff.R
Use different naming conventions within the same project
Use versions for code analysis_v1.2.R = analysis_final_v2_really_final_USE_THIS_ONE.R

Sometimes: include dates (e.g. data_2024_01_15.csv)

Naming conventions

i_use_snake_case                   # This is my preference

otherPeopleUseCamelCase            # This sometimes works

some.people.use.periods            # This is dangerous, especially in python

And_aFew.People_RENOUNCEconvention # You need help

ALL_CAPS                           # Reserved for super important stuff

Version control 101

What is Version Control?

A system that tracks changes to files over time

Think of it as:

A time machine for your code
A detailed history book of your project
A safety net for your work
A collaboration tool for teams

Git vs GitHub

Git

Software
Something you install in your computer
A command line tool (but we can bypass that)

GitHub

A service (offered by Microsoft)
Lives “on the cloud”
Provides a graphical interface
How you connect your computer to your colleague’s computer
Back up of your code

Git vs GitHub

Git

Git was created by Linus Torvalds (creator of Linux) in 2005

Key Features:

Free and open source: No licensing costs
Ample documentation
The most popular version control system

Four Git concepts

Repository

Clone

Commit

Push / Pull

Git Concepts: Repository

Repository (repo): A directory containing your project and its complete history

Think of it as:

A project folder with superpowers
A database of all your changes
A collaboration space for your team

Types:

Local repository: On your computer
Remote repository: On a server (like GitHub)

Example

The folder on my computer called tuna_data is the local repo
What you saw online was the remote repo
1 folder = 1 RStudio project = 1 repo

Git Concepts: Cloning

Cloning: The process of bringing a repo from the remote to local for the first time

When you clone a repo:

You get all files available in the remote, at that point in time
You only do this once per project (unless you have more than one computer)
RStudio has a nice way of doing it without having to use the command line
Cloning is just GitHub’s way of referring to “download and connect remote to local”

Git Concepts: Commit

Commit: A snapshot of your project at a specific point in time

Each commit contains:

Changes: What files were modified
Message: Description of what was changed and why
Author: Who made the changes
Timestamp: When the changes were made
Unique ID: A hash that identifies this commit

Think of commits as:

Checkpoints in your project’s history
Save points in a video game
Pages in a project’s storybook

Git Concepts: Push / Pull

Refers to uploading and downloading changes to and from the remote

With a push, you decide when you send a bundle of commits
With a pull, you decide when you bring down your colleague’s version of code

Git Workflow: The Basic Cycle

Modify files in your working directory
Stage changes you want to commit
Commit staged changes with a message
Repeat as you work on your project

Push the changes up to GitHub

This cycle creates a timeline of your project’s development

When / What to commit

No specific rule
Whatever a “change” might look like
Sometimes Added color to points in figure 2 is enough (one line in one file)
Sometimes Refactored code to work with new package is appropriate (many lines in many files)
Small commits and frequent commits are best

Writing Good Commit Messages

Good commit messages:

Are descriptive and specific
Explain WHAT and/or WHY and/or HOW things changed
Are concise but informative

Examples:

Add sea surface temperature data processing function
Fix bug in monthly trend calculation
Update README with installation instructions
stuff
fixed things

GitHub vs DropBox/Box/GoogleDrive…

Designed for the intended task
You have granular control over what is updated and when
Allows multiple versions of same project to evolve in parallel
- e.g. Your pipeline is done, mine isn’t yet

An example

mex_ports repository
- I created a new R script on August 1st
- Emma (student intern) worked on other files in the repo Aug 5 - Aug 11
- Emma was having trouble reading data I had created
- I had made a minor mistake, which I fixed
Note how I simply overwrite the file
No need to have V1 (didn’t work) and V2 (works) in the same repo
If I ever want to go back to an earlier version, I can!

Resources for Learning More

Git and GitHub:

R and Git:

Interesting reads

The Ultimate Guide to Git Branching Strategies

Next class

Hands-on: We’ll play with Git and GitHub

Install git and create a GitHub account

Instructions available here, and under Week 2 in the course’s website

Learning Objectives - Revisited

By the end of this week, you should be able to:

Understand the importance of reproducibility
Create the basic structure for an organized project
Explain the difference between Git and GitHub
Create and “clone” a “repo”
“commit” and “push”

Assignment

Assignment: Build Your Portfolio

Due Sunday, September 6, 2026 11:59 PM via Canvas:

Create a GitHub repository for your portfolio (I suggest you call it portfolio)

The repository should be public, or shared with us: jcvdav jreigrut

Add information to your README.md file
Make at least 2 meaningful commits
Share your repository link via Canvas

Grading criteria:

Repository:

50%: exists
30%: has a README.md file with content (name, about you, about the project)
10%: has a .gitignore file
10%: has at least two commits, in addition to the default one

Remember: Version control is a skill that gets better with practice. Start using Git for everything!

Extra slides

Essential Git Commands

We can do most things in RStudio, but you should know:

Getting started
- git init: Create a new repository
- git clone: Copy a repository from remote to local
Daily workflow
- git status: See what’s changed
- git add: Stage changes for commit
- git commit: Save changes with a message
Connecting with GitHub
- git push: Upload changes to remote repository
- git pull: Download changes from remote repository

GitHub Workflows

By yourself

Clone: Download the repository to your computer
Commit: Make and save your changes
Push: Upload your changes to GitHub

In a group

Clone: Download the repository to your computer
Branch: Create a separate line of development
Commit: Make and save your changes
Push: Upload your changes to GitHub
Pull Request: Propose changes to the original repository
Review: Get feedback and approval from collaborators
Merge: Integrate approved changes into the main project

References

Baker, Monya. 2016. “1,500 Scientists Lift the Lid on Reproducibility.” Nature 533 (May): 452–54. https://doi.org/10.1038/533452a.

Braga, Pedro Henrique Pereira, Katherine Hébert, Emma J Hudgins, et al. 2023. “Not Just for Programmers: How GitHub Can Accelerate Collaborative and Reproducible Research in Ecology and Evolution.” Methods Ecol. Evol. 14 (June): 1364–80.

Freedman, Leonard P, Iain M Cockburn, and Timothy S Simcoe. 2016. “The Economics of Reproducibility in Preclinical Research.” PLOS Biology 13 (June): e1002165. https://doi.org/10.1371/journal.pbio.1002165.

Lowndes, Julia S Stewart, Benjamin D Best, Courtney Scarborough, et al. 2017. “Our Path to Better Science in Less Time Using Open Data Science Tools.” Nature Ecology & Evolution 1 (6): 0160.

Keeping track of your code with Git and GitHub

Assigned Readings

Learning Objectives

Today’s Agenda

Version control and reproducibility

Why Should You Care?

The Problem: Research can be chaotic

Reproducible Research

The Reproducibility Crisis

Real-World Example: Climate Data Analysis

My drawer folder of shame

Organizing your projects

File vs path

Project Structure Fundamentals

Recommended Project Structure

Example

File Naming Best Practices

Naming conventions

Version control 101

What is Version Control?

Git vs GitHub

Git

GitHub

Git vs GitHub

Git

Four Git concepts

Git Concepts: Repository

Git Concepts: Cloning

Git Concepts: Commit

Git Concepts: Push / Pull

Git Workflow: The Basic Cycle

When / What to commit

Writing Good Commit Messages

GitHub: The Social Platform for Code

GitHub vs DropBox/Box/GoogleDrive…

An example

Resources for Learning More

Next class

Learning Objectives - Revisited

Assignment

Assignment: Build Your Portfolio

Extra slides

Essential Git Commands

GitHub Workflows

References