Vibe Coding ✨ The Art of Coding with Flow

---

layout: true
 
<div class="my-footer">

<a href="https://DataScience4Psych.github.io/DataScience4Psych/" target="_blank">Data Science for Psychologists</a>

</div>

---

# Hello world!

---

# Learning Goals

---

## Learning Goals

By the end of this session, you will be able to...

- Define vibe coding and explain how it enhances the coding experience
- Use AI tools (GitHub Copilot, ChatGPT, Claude) to write code with better flow and aesthetics
- Apply the five core principles of vibe coding to your R programming practice
- Leverage AI assistance to achieve flow state and improve code quality
- Evaluate when AI-assisted coding enhances vs. hinders the vibe

---

# What is Vibe Coding?

---

## Vibe Coding Defined

- ✨ **Aesthetic pleasure** in writing code

- 🌊 **Flow state** and intuitive development

- 🎨 **Elegance** over complexity

- 🎵 **Rhythm** in your coding process

---

## The Philosophy Behind Vibe Coding

.pull-left[
**Traditional Coding:**
- Focus on solving problems
- Efficiency first
- "Get it working"
- Technical correctness
]

.pull-right[
**Vibe Coding:**
- Focus on the experience
- Joy and flow first
- "Make it feel good"
- Aesthetic harmony
- *Then* optimize
]

---

# Core Principles of Vibe Coding

---

## Principle 1: Embrace the Flow 🌊

**How to achieve flow in coding:**

- Minimize distractions (notifications off 🔕)

- Choose tasks that match your skill level

- Set clear, achievable goals for each session

- Use tools that feel natural and responsive

.tip[
**Tip:** Set up your environment before you start. Good music 🎧, comfortable setup, and a clear plan help you enter flow faster.
]

---

## Principle 2: Code with Intention 🎯

**Bad vibe code:**

``` r
# Unclear, rushed code
x <- filter(data, a > 5)
y <- select(x, b, c, d)
z <- mutate(y, e = b * c)
```

**Good vibe code:**

``` r
# Clear, intentional pipeline
analysis_data <- raw_data %>%
 filter(age > 5) %>%
 select(participant_id, response_time, accuracy) %>%
 mutate(performance_score = response_time * accuracy)
```

---

## Principle 3: Aesthetics Matter 🎨

.pull-left[
**Elements of aesthetic code:**
- Consistent formatting
- Meaningful names
- Balanced spacing
- Visual alignment
- Color-coded syntax
]

.pull-right[
**Benefits:**
- Easier to read
- Fewer bugs
- More enjoyable
- Better collaboration
- Faster debugging
]

---

## Principle 4: Trust Your Intuition 🤔

**Listen to your gut:**

- Does this function name make sense? ❓

- Is this too complex for what it does? 😕

- Would I understand this tomorrow? 📆

- Does this feel elegant or clunky? ⚖️

---

## Principle 5: Create a Vibe-Friendly Environment 🏠

**Physical setup:**
- Comfortable chair and desk
- Good lighting (natural or warm)
- Plants or personal touches 🌰
- Clean, organized space

**Digital setup:**
- Pleasing color theme
- Organized file structure
- Helpful keyboard shortcuts
- Background music (if it helps)

---

# AI Tools for Vibe Coding

---

## AI as Your Vibe Coding Partner 🤖

**Popular AI coding assistants:**

.pull-left[
**GitHub Copilot**
- Inline code suggestions
- Context-aware completions
- Maintains your flow state
- Learns your coding style
]

.pull-right[
**ChatGPT / Claude**
- Code explanation & refactoring
- Debugging assistance
- Pattern suggestions
- Documentation help
]

---

## When AI Enhances the Vibe ✨

**Good vibe AI use:**

- ✅ **Autocomplete mundane code** - Let AI handle boilerplate while you focus on logic

- ✅ **Suggest better names** - "What's a clearer name for this function?"

- ✅ **Refactor for clarity** - "How can I make this more readable?"

- ✅ **Explain complex code** - Understanding beats copy-pasting

- ✅ **Generate test cases** - AI excels at creating thorough examples

---

## When AI Kills the Vibe ❌

**Bad vibe AI use:**

- ❌ **Accepting without understanding** - Code you can't explain isn't yours

- ❌ **Over-reliance on suggestions** - Kills your creative flow

- ❌ **Using outdated patterns** - AI may suggest old approaches

- ❌ **Copying complex logic blindly** - Recipe for future debugging pain

---

## Vibe Coding with GitHub Copilot

**Vibe-enhancing Copilot workflow:**
.midi[

``` r
# 1. Write a clear comment about your intent
# Calculate average response time by participant group

# 2. Let Copilot suggest, but review carefully
response_data %>%
  group_by(participant_group) %>%
  summarise(
    avg_response_time = mean(response_time_ms, na.rm = TRUE),
    n = n()
  )

# 3. Refine the suggestion to match your style
# 4. Keep coding with momentum!
```
]
--

---

## Vibe Coding with ChatGPT/Claude

**Example conversation:**

.pull-left[
**You:** 
"I have this code that works but feels clunky. How can I make it more elegant?"

``` r
for (i in 1:nrow(data)) {
 if (data$age[i] > 18) {
 data$adult[i] <- "yes"
 } else {
 data$adult[i] <- "no"
 }
}
```
]

``` r
data <- data %>%
 mutate(
 adult = if_else(
 age > 18, 
 "yes", 
 "no"
 )
 )
```
]

---

## The Vibe Coding AI Workflow ♻️

**The cycle:**

1. 💭 **Think** - Clarify your intent first

2. ⌨️ **Code** - Start writing with your own understanding

3. 🤖 **Assist** - Let AI suggest improvements or completions

4. 👀 **Review** - Does the AI suggestion feel right?

5. 🎨 **Refine** - Adapt to your style and standards

6. ✅ **Test** - Verify it works as expected

---

## AI for Code Review and Refactoring

**Ask AI to:**

- "Review this function for readability"
- "Suggest better variable names"
- "How can I make this more idiomatic in R?"
- "What edge cases am I missing?"
- "Refactor this to use tidyverse patterns"

**Example prompt:**
```
"I wrote this code for data cleaning. Can you suggest 
improvements for readability and maintainability while 
keeping the tidyverse style?"
```

---

## Prompt Engineering for Vibe Coding

**Vibe-friendly prompts:**

.pull-left[
**Specific:**
- "Write a function to calculate Z-scores that follows tidyverse style"
- "Refactor this loop into a purrr::map() call"
]

.pull-right[
**Context-aware:**
- "Using ggplot2, create an elegant visualization of..."
- "In the style of the tidyverse, how would you..."
]

**Include your constraints:**
```
"Create a readable function for [task] that:
- Uses descriptive variable names
- Includes error handling
- Follows tidyverse conventions
- Has a clear return value"
```

---

## Maintaining Your Vibe with AI

**Best practices:**

1. **Set boundaries** - Decide when to use AI vs. code independently

2. **Maintain ownership** - Understand every line AI suggests

3. **Preserve your style** - Adapt AI suggestions to your aesthetic

4. **Stay critical** - Question AI recommendations

5. **Learn continuously** - Use AI to discover new patterns, not replace learning

---

# Vibe Coding in R and Data Science

---

## The Tidyverse: Vibe Coding Champion 🏆

**Why the tidyverse has good vibes:**

``` r
# Reads like a sentence, flows naturally
starwars %>%
  filter(!is.na(height)) %>%
  group_by(species) %>%
  summarise(
    avg_height = mean(height),
    count = n()
  ) %>%
  arrange(desc(avg_height))
```

---

## Example: Data Cleaning with Good Vibes

``` r
d <- read.csv("data.csv")
d <- d[complete.cases(d),]
d$x2 <- d$x * d$x
d <- d[d$y > 0,]
names(d)[3] <- "z"
```
]

``` r
clean_data <- 
 read_csv("data.csv") %>%
 drop_na() %>%
 mutate(x_squared = x^2) %>%
 filter(response > 0) %>%
 rename(score = original_score)
```
]

---

## Visualization: Where Vibe Really Shines 🌈

--
.midi[

``` r
ggplot(data = penguins, 
       aes(x = flipper_length_mm, 
           y = body_mass_g, 
           color = species)) +
  geom_point(alpha = 0.7, size = 3) +
  scale_color_viridis_d() +
  theme_minimal(base_size = 14) +
  labs(
    title = "Palmer Penguins: Body Mass vs. Flipper Length",
    x = "Flipper Length (mm)",
    y = "Body Mass (g)",
    color = "Species"
  )
```
]
--

---

## Function Writing: The Vibe Way

--
.midi[
**Bad vibe function:**

``` r
f <- function(x, y, z) {
 a <- x + y
 b <- a * z
 c <- sqrt(b)
 return(c)
}
```
]

--
.midi[
**Good vibe function:**

``` r
calculate_performance_score <- function(accuracy, speed, weight = 1.0) {
 raw_score <- accuracy + speed
 weighted_score <- raw_score * weight
 normalized_score <- sqrt(weighted_score)
 
 return(normalized_score)
}
```
]
--

---

# Practical Vibe Coding Tips

---

## Tip 1: Use Consistent Style 🎨

**Popular R style guides:**
- [Tidyverse Style Guide](https://style.tidyverse.org/)
- Google's R Style Guide
- Your own consistent approach

---

# Key style elements:

``` r
# Good spacing
result <- function_name(argument1 = value1, argument2 = value2)

# Clear structure
if (condition) {
  action()
} else {
  alternative_action()
}

# Meaningful names
user_age <- calculate_age(birth_date)
```

---

## Tip 2: Comment with Purpose 📝

``` r
# Add 1 to x
x <- x + 1

# Loop through data
for (i in 1:10) {
  # Print i
  print(i)
}
```
]

``` r
# Adjust for zero-indexing
position <- position + 1

# Process each participant
# Note: Excluding participant 7 
# due to equipment failure
for (id in valid_participants) {
  analyze_responses(id)
}
```
]

---

## Tip 3: Take Breaks ☕

--
.pull-left[
**The Pomodoro Vibe:**
- Code for 25-50 minutes
- Take a 5-10 minute break
- Repeat with good energy

]
.pull-right[
**Benefits:**
- Maintains mental freshness
- Prevents burnout
- Keeps the creative flow going
- Better problem-solving
]
--

---

## Tip 4: Refactor with Joy ♻️

**Refactoring opportunities:**
- Repetitive code → Functions
- Nested conditionals → Guard clauses
- Long functions → Smaller, focused functions
- Cryptic names → Descriptive names

---

**The refactoring vibe:**

``` r
# Before: Repetitive and dull
mean_height <- mean(data$height)
mean_weight <- mean(data$weight)
mean_age <- mean(data$age)

# After: Elegant and satisfying
summary_stats <- data %>%
 summarise(across(c(height, weight, age), mean))
```

---

## Tip 5: Celebrate Small Wins 🎉

**Celebrate when you:**
- Write an elegant function ✅
- Fix a tricky bug 🐛
- Create a beautiful visualization 🎨
- Refactor messy code ✨
- Learn something new 📚

---

# Common Vibe Killers (and How to Fix Them)

---

## Vibe Killer 1: Analysis Paralysis 🤔

**The fix:**
- Start with a minimal viable solution
- Iterate and improve
- "Done is better than perfect" (then refine)

``` r
# Don't wait for the perfect approach
# Start simple:
preliminary_result <- data %>%
 filter(condition == "treatment") %>%
 summarise(mean_score = mean(score))

# Then enhance:
# Add error handling, visualization, etc.
```

---

## Vibe Killer 2: Error Frustration 😡

**The fix:**
- Errors are learning opportunities 📚
- Read error messages carefully
- Break the problem into smaller pieces
- Take a short walk if stuck

---

## Vibe Killer 3: Comparison 👀

**The fix:**
- Focus on your own progress
- Everyone starts somewhere
- Celebrate your unique coding style
- Learn from others, don't compete

---

## Vibe Killer 4: Messy Environment 🌪

**The fix:**

``` r
# Regular cleanup ritual
# 1. Archive old experiments
archive/
  old_analysis_v1.R
  abandoned_approach.R

# 2. Organize current work
scripts/
  01_data_cleaning.R
  02_exploratory_analysis.R
  03_modeling.R

# 3. Clear working directory
rm(list = ls()) # But save important objects first!
```

---

# Building Your Vibe Coding Practice

---

## Start Your Vibe Coding Journal 📓

**What to include:**
- Elegant code snippets you're proud of
- Problems solved and how
- New techniques learned
- Moments of flow and insight
- Personal coding preferences

**Benefits:**
- Reinforces learning
- Creates a personal reference
- Shows your progress
- Identifies patterns in your work

---

## Find Your Coding Rhythm 🎵

**Questions to explore:**
- What time of day do you feel most creative?
- Does music help or hinder? (If yes, what type?)
- Short bursts or long sessions?
- Morning freshness or evening focus?

**Experiment and adjust:**
- Try different schedules
- Test various environments
- Notice what feels best
- Honor your natural rhythms

---

## Build a Vibe Coding Community 👥

**Ways to connect:**
- Join R communities (R-Ladies, local R user groups)
- Share your code and visualizations
- Participate in TidyTuesday
- Attend coding meetups
- Pair program with friends

---

# Vibe Coding Challenges

---

## Your Turn: Vibe-ify This Code 🔧

Take this code and improve its vibe:
.pull-left[

``` r
d <- read.csv("data.csv")
d <- d[d$age > 18,]
d$bmi <- d$weight / (d$height^2)
m <- mean(d$bmi)
print(m)
```
]
.pull-right[
**Consider:**
- Variable names
- Readability 
- Flow
- Comments (if needed)
]
]
---

**Possible solution on next slide...**

---

## Vibe-ified Solution ✨

``` r
# Calculate BMI for adult participants
participant_data <- read_csv("data.csv") %>%
 filter(age > 18) %>%
 mutate(
 bmi = weight_kg / (height_m^2)
 )

# Summary statistics
average_bmi <- participant_data %>%
 summarise(mean_bmi = mean(bmi, na.rm = TRUE)) %>%
 pull(mean_bmi)

# Report findings
glue::glue("Average BMI for adult participants: {round(average_bmi, 2)}")
```

---

# Advanced Vibe Coding

---

## Vibe Coding with Projects 📁

**Example R project structure:**
```
my_analysis/
├── data/
│   ├── raw/
│   └── processed/
├── scripts/
│   ├── 01_import.R
│   ├── 02_clean.R
│   └── 03_analyze.R
├── outputs/
│   ├── figures/
│   └── tables/
├── R/
│   └── custom_functions.R
└── README.md
```

---

## Vibe Coding with Version Control 💾

**Bad commit messages:**
```
"fixed stuff"
"asdf"
"final version"
"FINAL VERSION 2"
```

**Good vibe commit messages:**
```
"Add data cleaning pipeline for survey responses"
"Improve visualization color scheme for accessibility"
"Refactor analysis functions for clarity"
"Document assumptions in modeling approach"
```

---

## Vibe Coding with Collaboration 🤝

**Best practices:**
- Write code others can understand
- Document your reasoning
- Be open to feedback
- Share your discoveries
- Celebrate team wins

**Code review mindset:**
- Assume positive intent
- Ask questions, don't criticize
- Offer suggestions, not demands
- Learn from each other

---

# The Future of Your Vibe Coding Journey

---

## Growing Your Vibe 🌱

**Continue to:**
- Experiment with new approaches
- Refine your personal style
- Stay curious and playful
- Share your knowledge
- Keep learning

**Remember:**
- There's no "perfect" vibe
- Your vibe will evolve
- What matters is enjoying the journey
- Code with intention and joy

---

## Resources for Vibe Coders 📚

**Learning resources:**
- [R for Data Science](https://r4ds.had.co.nz/) - The vibe-friendly intro
- [Tidyverse Style Guide](https://style.tidyverse.org/) - For consistent aesthetics
- [The Programmer's Brain](https://www.manning.com/books/the-programmers-brain) - Understanding flow
- [#TidyTuesday](https://github.com/rfordatascience/tidytuesday) - Weekly practice

**Community:**
- [RStudio Community](https://community.rstudio.com/)
- [R-Ladies](https://rladies.org/)
- [R4DS Online Learning Community](https://www.rfordatasci.com/)

---

# Summary: Learning Goals Achieved

---

## What We've Learned

Today, you should now be able to...

.pull-left[
### Concepts
- ✅ Define vibe coding philosophy
- ✅ Five core principles
- ✅ AI tools for vibe coding
- ✅ Vibe-friendly workflows
]

.pull-right[
### Skills
- ✅ Use GitHub Copilot effectively
- ✅ Prompt ChatGPT/Claude for better code
- ✅ Achieve flow state while coding
- ✅ Refactor with intention and aesthetics
- ✅ Evaluate AI suggestions critically
]

---

## Summary: The Vibe Coding Manifesto 📜

1. Code should be a joy to write and read
2. Aesthetics and clarity matter
3. Flow state enhances creativity
4. Small, intentional improvements compound
5. The journey is as important as the destination
6. Good vibes lead to better code
7. Community makes us all better

---

# Questions? ❓

---

## Contact Information

- Email: _garrissm@wfu.edu_
- Website: [smasongarrison.com](https://www.smasongarrison.com)
- GitHub: [github.com/smasongarrison](https://github.com/smasongarrison)