Data Visualization with Matplotlib - Discovery Challenge

Estimated Time: 60 minutes Course: AI-1 Data Analysis and Data Science Activity: 13

Master data visualization fundamentals! This activity teaches you to create professional line plots, bar charts, and scatter plots - essential skills for communicating data insights.

🎯 Learning Objectives

By the end of this activity, you will be able to:

Create line plots to visualize trends and time-series data
Build bar charts for categorical comparisons and distributions
Generate scatter plots to explore relationships between variables
Customize visualizations with labels, titles, colors, and legends
Combine multiple plots in a single figure using subplots

🚀 Getting Started (See Results in 30 Seconds!)

Choose your development environment based on your needs:

Option One: VS Code + Google Colab (Recommended for Developers)

Best for: Local development with cloud runtime consistency

Setup (first time only):

Open in VS Code:
bash
```
code activity-13-matplotlib.ipynb
```
Select Google Colab Kernel:
- Click "Select Kernel" (top-right) OR
- Press Cmd+Shift+P -> Type "Select Kernel" -> Choose "Google Colab"
Authenticate:
- Browser opens for Google login
- Grant permissions
- Return to VS Code (kernel connects automatically)
Run First Cell:
- Click the play button on any code cell
- Imports execute on Colab's cloud runtime
- Output appears directly in VS Code

Advantages:

✅ Runtime consistency with student environment (same library versions)
✅ Familiar IDE with extensions and shortcuts
✅ Git integration for version control
✅ Offline editing (execution requires internet)

First Connection: ~30-60 seconds Subsequent Connections: ~10-15 seconds

Option 2: Web Google Colab (Fallback for Students)

Best for: Quick access from any device, no setup required

Setup:

Open in Browser:
- Go to: https://colab.research.google.com
- Click "File" -> "Upload notebook"
- Select activity-13-matplotlib.ipynb
Run Cells:
- Click "Runtime" -> "Run all" to execute all cells
- Or run cells individually: Shift+Enter

Advantages:

✅ No installation needed (works on Chromebooks, tablets)
✅ Shareable via link
✅ Auto-save to Google Drive

Note: Web Colab has the same free tier limits as VS Code integration

Option 3: Local Jupyter (Quick Edits, No Cloud)

Best for: Offline work, quick edits, debugging

Setup:

bash

# Install dependencies (one-time)
pip install -r requirements.txt

# Start Jupyter Notebook
jupyter notebook activity-13-matplotlib.ipynb

# Or use Jupyter Lab (modern interface)
jupyter lab activity-13-matplotlib.ipynb

Advantages:

✅ Offline development
✅ Fastest iteration for quick changes

Limitations:

⚠️ Library version differences vs Colab (may see different plot rendering)
⚠️ Different environment than students will use

Use When: Quick edits, no internet, testing local changes

⚙️ Runtime Requirements

Execution Time

Colab (cloud): ~3-4 minutes
Local Jupyter: ~3-4 minutes

Memory Usage

Peak RAM: ~200MB
Colab Free Tier: 12GB available ✅ Sufficient

GPU Requirement

This Activity: ❌ Not needed (Matplotlib uses CPU for rendering)

Colab Tier Compatibility

Free Tier: ✅ Sufficient for this activity
Pro Tier: Not necessary

💡 Google Colab Free Tier Limits

Before starting, understand these session limits:

Limit	Value	Impact
Session Duration	12 hours max	Restart every 12 hours
Idle Timeout	90 minutes	Disconnects if inactive
RAM	12GB	Sufficient for this activity

💡 Tips:

Save frequently: Press Cmd+S (Mac) or Ctrl+S (Windows)
Avoid long breaks: Colab disconnects after 90 min idle
Re-run from top: After disconnection, run cells from beginning to restore state

🎯 What's Already Working

65% of the code is implemented for you:

✅ Imports configured (numpy, pandas, matplotlib ready to use)
✅ Sample data provided (datasets for practicing visualizations)
✅ Example plots included (see working visualizations immediately)
✅ Helper functions (data generation utilities)
⚠️ Line plot exercises (3 TODOs for you to complete)
⚠️ Bar chart exercises (3 TODOs for you to complete)
⚠️ Scatter plot exercises (2 TODOs for you to complete)

🔍 Explore the Working Code First

Before jumping into TODOs:

Run all cells to see example visualizations
Examine the plots generated by completed sections
Read the helper functions to understand available utilities
Review the plotting patterns to see best practices

📋 Tasks to Complete

TODO 1: Line Plot - Sales Trends (Easy)

Estimated Time: 10 minutes

What You'll Build: Create a line plot showing monthly sales trends over a year.

Success Criteria:

✅ Plot monthly sales data using plt.plot()
✅ Add x-axis label ("Month") and y-axis label ("Sales ($)")
✅ Add descriptive title ("Monthly Sales Trends 2024")
✅ Display plot with plt.show()

Hints:

💡 Use plt.plot(x_data, y_data) to create line plot
💡 Add labels: plt.xlabel("text"), plt.ylabel("text")
💡 Add title: plt.title("text")
💡 Always call plt.show() to display the plot

Expected Output:

diff

[Line plot showing upward sales trend from January to December]
- X-axis: Months (Jan-Dec)
- Y-axis: Sales values (0-10000)
- Blue line connecting data points

Key Concepts:

Line Plots: Best for showing trends over time or continuous data
Axis Labels: Essential for plot readability and context
Titles: Provide clear description of what the plot shows

Code Location: Cell 4 (search for # TODO 1:)

TODO 2: Line Plot - Multiple Lines (Medium)

Estimated Time: 12 minutes

What You'll Build: Create a line plot comparing sales trends for two products.

Success Criteria:

✅ Plot two lines on the same figure (Product A and Product B)
✅ Use different colors for each line
✅ Add legend to distinguish products
✅ Customize line styles (solid vs dashed)

Hints:

💡 Plot multiple lines: plt.plot(x, y1, label="Product A") then plt.plot(x, y2, label="Product B")
💡 Set colors: color='blue' or color='red'
💡 Set line styles: linestyle='-' (solid) or linestyle='--' (dashed)
💡 Display legend: plt.legend()

Expected Output:

less

[Line plot with two lines]
- Blue solid line: Product A sales
- Red dashed line: Product B sales
- Legend in corner identifying each product

Key Concepts:

Multiple Series: Comparing datasets on the same plot
Color Coding: Using colors to distinguish data series
Legends: Labeling plot elements for clarity

Code Location: Cell 6 (search for # TODO 2:)

TODO 3: Line Plot - Advanced Styling (Medium)

Estimated Time: 10 minutes

What You'll Build: Create a styled line plot with markers, grid, and custom figure size.

Success Criteria:

✅ Add markers to data points (circles, squares, etc.)
✅ Enable grid for easier value reading
✅ Set custom figure size (10x6 inches)
✅ Add marker size and line width customization

Hints:

💡 Add markers: marker='o' (circle), marker='s' (square)
💡 Enable grid: plt.grid(True)
💡 Set figure size: plt.figure(figsize=(10, 6)) before plotting
💡 Customize: linewidth=2, markersize=8

Expected Output:

css

[Line plot with enhanced styling]
- Data points marked with circles
- Grid lines visible
- Larger figure size (10x6)
- Thicker line for visibility

Key Concepts:

Markers: Highlight individual data points
Grids: Improve readability of values
Figure Size: Control plot dimensions for presentations

Code Location: Cell 8 (search for # TODO 3:)

TODO 4: Bar Chart - Category Comparison (Easy)

Estimated Time: 10 minutes

What You'll Build: Create a bar chart comparing sales across different product categories.

Success Criteria:

✅ Create bar chart using plt.bar()
✅ Display 5 product categories on x-axis
✅ Show sales values on y-axis
✅ Add value labels on top of bars

Hints:

💡 Basic bar chart: plt.bar(categories, values)
💡 Add value labels: Loop through bars and use plt.text(x, y, value)
💡 Rotate x-axis labels if needed: plt.xticks(rotation=45)
💡 Set custom bar colors: color=['blue', 'green', 'red', ...]

Expected Output:

vbnet

[Bar chart with 5 categories]
- Categories: Electronics, Clothing, Food, Books, Toys
- Different colored bars for each category
- Value labels displayed on top of each bar

Key Concepts:

Bar Charts: Best for comparing discrete categories
Value Labels: Showing exact numbers for precision
Color Coding: Highlighting different categories

Code Location: Cell 10 (search for # TODO 4:)

TODO 5: Bar Chart - Grouped Bars (Medium)

Estimated Time: 15 minutes

What You'll Build: Create a grouped bar chart comparing Q1 and Q2 sales for multiple products.

Success Criteria:

✅ Display two bars per category (Q1 and Q2)
✅ Position bars side-by-side using offset
✅ Use different colors for each quarter
✅ Add legend to identify quarters

Hints:

💡 Calculate bar positions: x = np.arange(len(categories))
💡 Offset bars: plt.bar(x - width/2, q1_values) and plt.bar(x + width/2, q2_values)
💡 Set bar width: width = 0.35
💡 Fix x-axis labels: plt.xticks(x, categories)

Expected Output:

diff

[Grouped bar chart]
- Two bars per product (Q1 in blue, Q2 in orange)
- Bars positioned side-by-side
- Legend distinguishing quarters

Key Concepts:

Grouped Bars: Comparing multiple values per category
Bar Positioning: Using numpy for precise placement
Width Control: Adjusting bar width for readability

Code Location: Cell 12 (search for # TODO 5:)

TODO 6: Bar Chart - Horizontal Bars (Easy)

Estimated Time: 8 minutes

What You'll Build: Create a horizontal bar chart for ranking data (top 5 products).

Success Criteria:

✅ Use plt.barh() for horizontal bars
✅ Sort data from highest to lowest
✅ Add value labels at end of bars
✅ Reverse y-axis so highest value is at top

Hints:

💡 Horizontal bars: plt.barh(categories, values)
💡 Sort data: Use pandas .sort_values(ascending=False)
💡 Reverse y-axis: plt.gca().invert_yaxis()
💡 Value labels: plt.text(value, y_position, f'{value}')

Expected Output:

css

[Horizontal bar chart]
- Bars extend from left to right
- Highest value at top
- Value labels at end of each bar

Key Concepts:

Horizontal Bars: Better for long category names or rankings
Sorting: Presenting data in meaningful order
Axis Control: Inverting axes for better readability

Code Location: Cell 14 (search for # TODO 6:)

TODO 7: Scatter Plot - Correlation Analysis (Medium)

Estimated Time: 10 minutes

What You'll Build: Create a scatter plot exploring the relationship between advertising spend and sales.

Success Criteria:

✅ Plot scatter points using plt.scatter()
✅ Add trend line (linear regression)
✅ Color points by region (use color map)
✅ Add colorbar to show region mapping

Hints:

💡 Basic scatter: plt.scatter(x_data, y_data)
💡 Color by category: c=region_codes, cmap='viridis'
💡 Add colorbar: plt.colorbar(label='Region')
💡 Trend line: Use np.polyfit() and np.poly1d()

Expected Output:

diff

[Scatter plot showing positive correlation]
- Points scattered showing trend
- Points colored by region
- Trend line overlaid
- Colorbar on right side

Key Concepts:

Scatter Plots: Visualizing relationships between variables
Correlation: Identifying trends in data
Color Mapping: Adding third dimension to 2D plot

Code Location: Cell 16 (search for # TODO 7:)

TODO 8: Scatter Plot - Subplots (Hard)

Estimated Time: 15 minutes

What You'll Build: Create a figure with 2x2 subplots showing different scatter plot analyses.

Success Criteria:

✅ Create 4 subplots using plt.subplots(2, 2)
✅ Plot different variable pairs in each subplot
✅ Customize each subplot independently
✅ Add overall figure title

Hints:

💡 Create subplots: fig, axes = plt.subplots(2, 2, figsize=(12, 10))
💡 Access subplot: axes[0, 0].scatter(x, y)
💡 Set subplot titles: axes[0, 0].set_title('Title')
💡 Overall title: fig.suptitle('Main Title', fontsize=16)
💡 Adjust spacing: plt.tight_layout()

Expected Output:

sql

[2x2 grid of scatter plots]
- Top-left: Price vs Sales
- Top-right: Marketing vs Sales
- Bottom-left: Quantity vs Revenue
- Bottom-right: Discount vs Profit
Each with appropriate labels and styling

Key Concepts:

Subplots: Comparing multiple visualizations in one figure
Figure Management: Controlling layout and spacing
Multi-variable Analysis: Exploring different relationships

Code Location: Cell 18 (search for # TODO 8:)

🧪 Testing Your Work

Manual Testing Checklist

Run All Cells:

All cells execute without errors
Outputs display correctly (no blank plots)
Visualizations match expected results
No warnings or deprecation messages

Visualization Validation:

All plots have titles
Axis labels are present and descriptive
Legends display when multiple series present
Colors are distinguishable
Text is readable (not overlapping)
Plot sizes are appropriate

Data Accuracy:

Plotted values match source data
Scales are appropriate (not truncated or distorted)
Trends are clearly visible

🛠 Helper Functions Already Included

You don't need to write these - they're already working! Just call them with your data:

Data Generation

python

generate_sales_data()        # Creates sample sales dataset
generate_scatter_data()      # Creates correlated x-y data
create_category_data()       # Generates categorical data

Utilities

python

add_value_labels(ax, bars)   # Adds labels on top of bars
format_currency(value)       # Formats numbers as currency
save_plot(filename)          # Saves figure to file

🔧 Troubleshooting

Common Issues & Solutions

Issue: "ModuleNotFoundError: No module named 'matplotlib'" Solution:

python

# In Colab, matplotlib is pre-installed
# If using local Jupyter, install dependencies:
pip install matplotlib

# Then restart kernel: Kernel → Restart

Issue: "Figure not displaying" or "Blank output" Solution:

Jupyter/Colab: Ensure %matplotlib inline magic is in first cell
Missing plt.show(): Add plt.show() at end of plotting code
Backend issues: Try plt.switch_backend('agg') then plt.show()
Clear figure: If rerunning cell, add plt.clf() before plotting

Issue: "Session disconnected" or "Kernel restarting" Solution:

Colab disconnects after 90 minutes idle
Action: Re-run cells from top to restore state
Prevention: Save frequently (Cmd+S / Ctrl+S)
Matplotlib plots regenerate quickly (no training needed)

Issue: "Text overlapping" or "Labels cut off" Solution:

python

# Fix overlapping x-axis labels
plt.xticks(rotation=45, ha='right')

# Adjust layout to prevent cutoff
plt.tight_layout()

# Manually adjust margins
plt.subplots_adjust(bottom=0.15, left=0.15)

# Increase figure size
plt.figure(figsize=(12, 8))

Issue: "Plot axis limits too tight" or "Data cut off" Solution:

python

# Set custom axis limits
plt.xlim(0, 100)  # x-axis from 0 to 100
plt.ylim(0, 500)  # y-axis from 0 to 500

# Add padding automatically
plt.margins(0.1)  # 10% padding on all sides

# Let matplotlib auto-scale
plt.autoscale()

Issue: "Colors don't match" or "Plot looks different than expected" Solution:

Matplotlib default colors may vary by version
Use explicit colors: color='blue' instead of relying on defaults
Use color codes: color='#1f77b4' for exact hex colors
Check colormap: Different versions may have different defaults

Issue: "Memory errors with large datasets" Solution:

python

# Reduce number of points plotted
# For scatter plots with 100k+ points:
sample_size = 10000
sample_indices = np.random.choice(len(data), sample_size, replace=False)
plt.scatter(data[sample_indices]['x'], data[sample_indices]['y'])

# Use rasterization for large plots
plt.scatter(x, y, rasterized=True)

# Clear figure after saving
plt.savefig('plot.png')
plt.close()  # Frees memory

Debugging Tips

Check Data Shape:

python

print(f"X data shape: {x_data.shape}")
print(f"Y data shape: {y_data.shape}")
# Should have compatible dimensions

Verify Plot Object:

python

# Check if plot was created
ax = plt.gca()  # Get current axis
print(f"Number of lines: {len(ax.lines)}")
print(f"Number of patches (bars): {len(ax.patches)}")

Inspect Data Range:

python

print(f"X range: {x_data.min()} to {x_data.max()}")
print(f"Y range: {y_data.min()} to {y_data.max()}")
# Ensure data is within expected range

Test Incrementally:
- Start with basic plot: plt.plot(x, y)
- Add styling one element at a time
- Verify each addition before proceeding
Use Matplotlib Documentation:
- Gallery for examples: https://matplotlib.org/stable/gallery/
- Function reference: https://matplotlib.org/stable/api/

📚 Library Versions

This activity was tested with Google Colab runtime as of November 2024:

Library	Version (Colab)	Notes
NumPy	One.26.4	Pre-installed
Pandas	2.1.4	Pre-installed
Matplotlib	3.8.0	Pre-installed

Note: Colab libraries update periodically. If you encounter version conflicts:

Check current versions: !pip list | grep -E "numpy|pandas|matplotlib"
This activity uses stable matplotlib API (minimal version differences)
Report issues to instructor if compatibility problems persist

🚀 Extension Challenges

Completed the main activity? Try these bonus challenges:

Beginner Extensions

Challenge 1: Create a pie chart showing market share distribution [+15 min]
Challenge 2: Build a histogram showing age distribution in a dataset [+15 min]
Challenge 3: Create a box plot comparing sales across regions [+20 min]

Intermediate Extensions

Challenge 4: Design an infographic-style dashboard with 4+ plot types [+45 min]
Challenge 5: Create animated plots showing data evolution over time [+1 hour]
Challenge 6: Build interactive plots with hover tooltips (using mpld3) [+1 hour]

Advanced Extensions

Challenge 7: Create publication-quality plots with custom styles and LaTeX labels [+1.5 hours]
Challenge 8: Build a complete data story with 6+ coordinated visualizations [+2 hours]

🏆 Success Criteria

Your activity is complete when:

✅ All 8 TODO exercises completed
✅ All cells execute without errors
✅ Plots display correctly with proper styling
✅ All visualizations have titles and labels
✅ Code follows matplotlib best practices
✅ Extension challenges attempted (optional but encouraged)

📤 Submission

Complete all TODOs in the notebook
Verify all cells run without errors (Runtime -> Run all, or Kernel -> Restart & Run All)
Check visualizations:
- All plots display correctly
- No blank outputs or cut-off labels
- Colors and legends are clear
Save your notebook:
- VS Code: Cmd+S / Ctrl+S
- Web Colab: File -> Save
- Local Jupyter: File -> Save and Checkpoint
Download completed notebook (if using web Colab):
- File -> Download -> Download .ipynb
Submit via course portal (follow course-specific instructions)

📚 Additional Resources

Matplotlib Resources

Matplotlib Gallery: https://matplotlib.org/stable/gallery/ - Browse examples by plot type
Matplotlib Tutorials: https://matplotlib.org/stable/tutorials/ - Official tutorials
Matplotlib Cheatsheet:
https://github.com/matplotlib/cheatsheets
matplotlib/cheatsheets
View on GitHub
- Quick reference
Seaborn Gallery: https://seaborn.pydata.org/examples/ - Higher-level plotting library

Visualization Concepts

Data Viz Catalog: https://datavizcatalogue.com/ - Choose right chart type
Color Brewer: https://colorbrewer2.org/ - Color scheme advice
Storytelling with Data: https://www.storytellingwithdata.com/ - Visualization best practices

Documentation

NumPy: https://numpy.org/doc/
Pandas: https://pandas.pydata.org/docs/
Google Colab Guide: https://colab.research.google.com/notebooks/intro.ipynb

Previous: Activity 12 - Pandas Data Manipulation (prerequisite for data preparation)
Next: Activity 14 - Advanced Visualizations (builds on these fundamentals)

🎉 Congratulations!

Once you complete this activity, you'll have:

✅ Mastered line plots for trend analysis
✅ Created bar charts for categorical comparisons
✅ Built scatter plots for correlation analysis
✅ Customized visualizations with professional styling
✅ Combined multiple plots using subplots

This activity equips you with essential data visualization skills used by data analysts and data scientists worldwide!

💬 Need Help?

Stuck on a TODO?

Re-read the hints section
Check the helper functions for available utilities
Review the matplotlib gallery for similar examples
Examine the example plots in earlier cells

Still stuck after 15 minutes?

Post in the course discussion forum
Ask during office hours
Email instructor with:
- Specific TODO number
- Error message screenshot or blank output
- What you've tried so far

Pro Tips:

💡 Always call plt.show() at the end of plotting code (not needed in some notebook environments, but good practice)
💡 Use plt.figure() before each new plot to avoid overlapping
💡 Check data types: Matplotlib expects numeric data for plotting
💡 Experiment with examples: Modify provided code to understand parameters
💡 Save your work: Use plt.savefig('plot.png', dpi=300, bbox_inches='tight') to export plots

Remember: Visualization is both art and science. It's OK to:

❌ Create "ugly" plots at first (styling improves with practice)
❌ Need multiple attempts to get layout right (even experts iterate)
❌ Ask for feedback on design choices (visualization is subjective!)

Happy plotting! 🚀📊✨

This activity follows the 65-70% implementation methodology: example visualizations provided, students complete targeted plotting exercises

Template 13: Matplotlib

📦 Project Files Included:

https://github.com/matplotlib/cheatsheets