IntroDataScience

Copilot-Assisted Coding

Learn to code with AI as your pair programming partner.

1. What is GitHub Copilot?

GitHub Copilot is an AI coding assistant that helps you write code faster and learn as you go.

Key Features

Code Completion đź’ˇ

Code Explanation đź“–

Code Generation 🤖

Chat Interface đź’¬

2. Setting Up Copilot in VS Code

Install the Extension

  1. Open VS Code
  2. Click Extensions (Ctrl/Cmd + Shift + X)
  3. Search for “GitHub Copilot”
  4. Click Install
  5. Sign in with your GitHub account

Get Access

Free for:

Paid for:

Trial:

Verify It’s Working

  1. Create a new Python file: test.py
  2. Type: def calculate_average(
  3. Copilot should suggest code!
  4. Press Tab to accept

3. How Copilot Works

The Technology

What Copilot Can Do

âś… Excellent at:

❌ Be careful with:

Privacy

4. Effective Prompting

The key to great Copilot suggestions is writing clear comments!

Good Comments

Be Specific:

# Bad: Calculate something
# Good: Calculate the average test score for students with > 85% attendance

def calculate_average_high_attendance(df):
    # Copilot suggests accurate code!

Include Context:

# Load student data from CSV, filter for age > 15,
# and return the top 10 students by test score

import polars as pl

def get_top_students(filepath):
    # Copilot understands what you want!

Use Examples:

# Convert temperature from Celsius to Fahrenheit
# Example: 0°C = 32°F, 100°C = 212°F

def celsius_to_fahrenheit(celsius):
    # Copilot generates: return (celsius * 9/5) + 32

Descriptive Function Names

# Good: Copilot understands intent
def filter_high_scoring_students(df, threshold=80):
    # Suggestion: return df.filter(pl.col("test_score") > threshold)

# Less clear
def filter_students(df, x):
    # Copilot is less certain what to do

Break Down Problems

# Instead of one big function, break it down:

# Step 1: Load the sales data from JSON
sales = pl.read_json("data/raw/sales.json")

# Step 2: Filter for transactions in 2024
sales_2024 = sales.filter(pl.col("date").str.starts_with("2024"))

# Step 3: Calculate total sales by category
category_totals = sales_2024.group_by("product_category").agg(
    pl.sum("total_amount").alias("total")
)

# Each step gives Copilot clear context!

5. Copilot Features

Inline Suggestions (Tab to Accept)

As you type, Copilot suggests completions:

# Type this:
def calculate_

# Copilot suggests:
def calculate_average(numbers):
    return sum(numbers) / len(numbers)

# Press Tab to accept
# Press Esc to reject

Multiple Suggestions:

Copilot Chat

Ask questions in natural language!

Open Chat:

Example Questions:

"How do I read a CSV file with Polars?"
"Explain this code: df.group_by('category').agg(pl.mean('value'))"
"What's wrong with this function?"
"How do I create a bar chart with Plotly?"

Explain Code

Right-click on code → “Copilot: Explain This”

# Select this code and ask Copilot to explain:
result = (
    df.filter(pl.col("age") > 18)
    .group_by("category")
    .agg(pl.mean("score"))
    .sort("score", descending=True)
)

# Copilot explains: "This code filters the DataFrame for rows where
# age is greater than 18, groups by category, calculates the mean
# score for each category, and sorts by score in descending order."

Fix Errors

When you have an error:

  1. Right-click on the error
  2. Select “Copilot: Fix This”
  3. Copilot suggests a fix!
# Error: NameError: name 'polars' is not defined
result = polars.read_csv("data.csv")

# Copilot suggests: Add import statement
import polars as pl
result = pl.read_csv("data.csv")

Generate Tests

def calculate_average(numbers):
    return sum(numbers) / len(numbers)

# Type: "test for calculate_average"
# Copilot generates:

def test_calculate_average():
    assert calculate_average([1, 2, 3, 4, 5]) == 3
    assert calculate_average([10, 20, 30]) == 20
    assert calculate_average([5]) == 5

6. Best Practices

Review All Suggestions âś…

Always check:

Example:

# Copilot suggests:
def divide_numbers(a, b):
    return a / b

# You should add:
def divide_numbers(a, b):
    if b == 0:
        raise ValueError("Cannot divide by zero")
    return a / b

Learn from the Code đź“š

Don’t just accept suggestions—understand them!

Good workflow:

  1. Accept Copilot’s suggestion
  2. Read through the code
  3. Ask Copilot Chat: “Explain this code”
  4. Try modifying it yourself
  5. Run it to see it work

Example:

# Copilot suggests:
df_sorted = df.sort("test_score", descending=True)

# Ask yourself:
# - What does descending=True do?
# - What happens to null values?
# - Can I sort by multiple columns?

# Experiment:
df_sorted = df.sort(["grade_level", "test_score"], descending=[False, True])

Don’t Blindly Accept 🚫

Be cautious when:

Example:

# Copilot might suggest:
def calculate_gpa(grades):
    return sum(grades) / len(grades)

# But you know GPAs use weighted averages with credits!
def calculate_gpa(grades, credits):
    total_points = sum(g * c for g, c in zip(grades, credits))
    total_credits = sum(credits)
    return total_points / total_credits

Use as a Learning Tool 🎓

Learning workflow:

  1. Try yourself first - Attempt the problem
  2. Get stuck? - Ask Copilot for help
  3. Compare - How does Copilot’s solution differ?
  4. Understand - Why is Copilot’s approach better/different?
  5. Practice - Try a similar problem yourself

Example:

# Your attempt:
students_over_80 = []
for row in df.iter_rows():
    if row['test_score'] > 80:
        students_over_80.append(row)

# Copilot's suggestion:
students_over_80 = df.filter(pl.col("test_score") > 80)

# Learning: Polars has built-in filtering! Much cleaner.

7. Example Workflow: Building a Data Analysis with Copilot

Let’s walk through a real analysis using Copilot:

Step 1: Define the Goal

# Analyze sales data to find:
# 1. Total sales by product category
# 2. Average transaction value by region
# 3. Top 5 best-selling products

import polars as pl
import plotly.express as px

Step 2: Load Data (with Copilot)

# Load sales data from JSON file
sales = pl.read_json("data/raw/sales.json")

# View the first few rows to understand structure
sales.head()

Copilot helps with the exact syntax!

Step 3: Calculate Metrics (with Copilot)

# Calculate total sales by product category
category_sales = sales.group_by("product_category").agg(
    pl.sum("total_amount").alias("total_sales")
).sort("total_sales", descending=True)

# Calculate average transaction value by region
region_avg = sales.group_by("region").agg(
    pl.mean("total_amount").alias("avg_transaction")
)

# Find top 5 best-selling products by quantity
top_products = sales.group_by("product_name").agg(
    pl.sum("quantity").alias("total_quantity")
).sort("total_quantity", descending=True).head(5)

Copilot suggests aggregations based on your comments!

Step 4: Visualize (with Copilot)

# Create bar chart of sales by category
fig_category = px.bar(
    category_sales,
    x="product_category",
    y="total_sales",
    title="Total Sales by Product Category"
)
fig_category.show()

# Create bar chart of average transaction by region
fig_region = px.bar(
    region_avg,
    x="region",
    y="avg_transaction",
    title="Average Transaction Value by Region"
)
fig_region.show()

Copilot knows Plotly syntax!

Step 5: Review and Refine

# You notice: Need to handle edge cases
# Ask Copilot Chat: "How do I filter out null values in Polars?"

# Copilot suggests:
sales_clean = sales.drop_nulls(subset=["total_amount"])

# Now re-run your analysis with clean data!

8. When to Use Copilot vs. Think Independently

Use Copilot When:

âś… Learning new syntax

# "How do I join two DataFrames in Polars?"
# Copilot shows you the syntax

âś… Writing boilerplate

# "Create a function to load and validate CSV data"
# Copilot generates the structure

âś… Stuck on a problem

# You've been stuck for 10 minutes
# Ask Copilot for ideas

âś… Exploring alternatives

# "What are different ways to filter this data?"
# Copilot shows options

Think Independently When:

🧠 Learning fundamentals

🧠 Understanding concepts

🧠 Problem-solving

🧠 Domain expertise

9. Ethical Considerations

Understanding Your Code

You are responsible for:

Always:

Academic Integrity

For students:

In this course:

Copilot Etiquette

Good practices:

10. Troubleshooting Copilot

Suggestions Not Appearing

Solutions:

  1. Check Copilot is enabled (bottom status bar)
  2. Sign in to GitHub
  3. Check subscription is active
  4. Reload VS Code
  5. Check extension settings

Suggestions Are Wrong

Solutions:

  1. Provide more context in comments
  2. Be more specific
  3. Try rephrasing
  4. Ask Copilot Chat for explanation
  5. Use your judgment!

Copilot Is Slow

Solutions:

  1. Check internet connection
  2. Reduce open files
  3. Restart VS Code
  4. Check GitHub status

Keyboard Shortcuts

Action Shortcut
Accept suggestion Tab
Reject suggestion Esc
Next suggestion Alt/Opt + ]
Previous suggestion Alt/Opt + [
Open Copilot Chat Ctrl/Cmd + Shift + I
Inline Chat Ctrl/Cmd + I

Example Conversations with Copilot Chat

Question 1: Learning Syntax

You: "How do I create a line chart with Plotly Express showing temperature over time?"

Copilot: "Here's how to create a line chart with Plotly Express:

import plotly.express as px
import polars as pl

df = pl.read_csv("weather.csv")
fig = px.line(df, x="date", y="temperature", title="Temperature Over Time")
fig.show()

The px.line() function creates a line chart. The x parameter is your time axis,
y is the value to plot, and title adds a chart title."

Question 2: Debugging

You: "I'm getting 'AttributeError: 'DataFrame' object has no attribute 'groupby''. What's wrong?"

Copilot: "In Polars, the method is group_by() with an underscore, not groupby().

Change:
df.groupby("category")

To:
df.group_by("category")

This is different from Pandas which uses groupby() without an underscore."

Question 3: Understanding Code

You: "Explain what this does:
df.with_columns((pl.col("price") * 1.1).alias("price_with_tax"))"

Copilot: "This code adds a new column called 'price_with_tax' to the DataFrame.

Breaking it down:
- pl.col("price") - references the price column
- * 1.1 - multiplies by 1.1 (adds 10%)
- .alias("price_with_tax") - names the new column
- with_columns() - adds this as a new column

The original DataFrame isn't modified; it returns a new DataFrame with the extra column."

Practice Exercise

Try using Copilot to solve this problem:

Task: Load the students dataset, filter for students with test scores above 80, group by grade level, and calculate the average attendance rate for each grade. Then create a bar chart.

Workflow:

  1. Write comments describing each step
  2. Let Copilot suggest code
  3. Review and understand each suggestion
  4. Run the code
  5. Debug if needed with Copilot’s help

Solution approach:

# Load students dataset from CSV
import polars as pl
import plotly.express as px

students = pl.read_csv("data/raw/students.csv")

# Filter for students with test scores above 80
high_scorers = students.filter(pl.col("test_score") > 80)

# Group by grade level and calculate average attendance
attendance_by_grade = high_scorers.group_by("grade_level").agg(
    pl.mean("attendance_rate").alias("avg_attendance")
)

# Create bar chart
fig = px.bar(
    attendance_by_grade,
    x="grade_level",
    y="avg_attendance",
    title="Average Attendance for High Scorers by Grade"
)
fig.show()

Next Steps

Official Resources:

Copilot is your AI pair programmer—use it to learn faster and code better! 🤖✨