AI-Assisted Programming Pedagogy: Workshop Guide

Social Media Analysis Workshop

Workshop Overview

Building on Your Experience: Combinatorial text generation → Social media data analysis
Focus: AI pedagogy strategies for programming instruction
Time: 10:40 AM - 12:30 PM

Three-Tier AI Pedagogical Framework

Level 1: Code Comprehension & Debugging

For students building foundational understanding

Prompt Examples:

“Explain this error message and how to fix it”
“Walk me through what this code block accomplishes step by step”
“I used str.split() before. How is str.split().str.len() different?”

Teaching Goal: Build confidence through understanding existing code

Level 2: Conceptual Application & Adaptation

For students connecting programming to disciplinary knowledge

Prompt Examples:

“How would I modify this approach for historical data?”
“In my text project, I used find/replace. How could I use similar pattern matching here?”
“What’s the computational thinking behind this solution?”

Teaching Goal: Connect technical skills to research questions

Level 3: Critical Evaluation & Extension

For students ready to evaluate computational approaches

Prompt Examples:

“What are the limitations of this algorithmic approach?”
“Compare the ethical considerations between generating creative text and analyzing real social media posts”
“How does this method compare to traditional research practices?”

Teaching Goal: Develop critical digital literacy

Session 1: Building Real Corpus with Bluesky API

Jupyter Notebook Workflow Management

Cell Organization Strategy:

Setup Cell: Libraries and authentication (run once)
Data Collection Cell: API calls (modify and re-run)
Processing Cell: Clean and structure data
Analysis Cells: Individual analyses (iterate with AI)
Visualization Cell: Final outputs

AI-Assisted Iteration Pattern:

Keep working code in separate cells
Use AI to modify copies before replacing originals
Save successful versions before experimenting

Getting Started with Real Data Collection

AI Prompts for Setup:

“Help me install the required libraries for Bluesky API analysis: atproto, pandas, matplotlib, seaborn”
“Show me how to securely authenticate with the Bluesky API using Google Colab secrets”
“I’m getting an authentication error with the AT Protocol. What might be wrong?”

AI Prompts for Data Collection:

“Write a function to collect posts from a specific Bluesky user handle”
“How do I search for posts containing specific hashtags on Bluesky?”
“Create a function that collects posts from multiple users and combines them into one dataset”
“My data collection is only getting 10 posts per user. How can I get more historical data?”

AI Prompts for Data Processing:

“Convert this Bluesky API response into a pandas DataFrame with relevant features”
“Help me extract engagement metrics, timestamps, and text features from Bluesky posts”
“I have inconsistent API response structures. How do I handle missing fields robustly?”

AI Iteration Strategy: Use prompts like:

“This code works but is slow. How can I optimize the data collection?”
“I’m getting rate limited by the API. What’s the best way to handle this?”
“How do I save my collected data as a CSV file for backup?”

Session 2: AI-Assisted Analysis with Real Data

Jupyter Cell Management for Iterative Analysis

Best Practices:

Test in new cells before modifying working code
Comment out previous versions instead of deleting
Use cell markdown to document AI conversations
Save successful iterations before experimenting further

Content Analysis on Your Corpus

AI Prompts for Pattern Recognition:

“Create a function to categorize social media posts as academic, literary, or general based on keywords”
“How do I analyze word frequency patterns in my social media corpus?”
“Build a function that identifies posts mentioning specific academic or cultural terms”

AI Prompts for Feature Extraction:

“Calculate engagement rates and text statistics for each post in my dataset”
“How do I identify posts with hashtags, mentions, or media attachments?”
“Create features that measure post complexity, sentiment, or topic relevance”

AI Debugging Prompts:

“My categorization function is putting everything in ‘general’. How do I debug this?”
“The regex patterns aren’t catching word variations like ‘researching’. Help me improve them”
“I want to add sentiment analysis to this content analysis. How would I modify the approach?”

Temporal Analysis with Your Data

AI Prompts for Time-Based Analysis:

“Analyze posting patterns by hour and day of week in my dataset”
“How do I identify unusual spikes or patterns in daily posting activity?”
“Calculate engagement differences between weekend and weekday posts”
“Create a function that finds the most active time periods in my corpus”

AI Iteration Examples:

“I want to analyze posting patterns around specific events. How do I identify date ranges with unusual activity?”
“My temporal analysis shows all posts from one day. Is my data collection working correctly?”
“How can I compare posting patterns between different authors in my corpus?”

Session 3: Visualization and Analysis Iteration

Jupyter Workflow for AI-Assisted Visualization

Cell Strategy for Visual Analysis:

Basic plot cell - get something working first
AI improvement cell - iterate with prompts
Final visualization cell - polished version
Interpretation cell - analysis of patterns

Creating Visualizations from Your Data

AI Prompts for Basic Visualization:

“Create a bar chart showing the distribution of content categories in my dataset”
“Make a scatter plot of word count vs engagement for my social media posts”
“Generate a timeline showing daily posting activity over the data collection period”
“Create a simple heatmap of posting activity by hour and day of week”

AI Prompts for Enhanced Visualization:

“Improve this basic bar chart with better colors, labels, and professional formatting”
“Create a multi-panel dashboard showing key patterns in my social media data”
“This scatter plot is too crowded. How can I make it clearer and more informative?”
“Add trend lines and statistical annotations to show correlations in my data”

AI Prompts for Specialized Analysis:

“Create a network visualization showing user mentions and interactions”
“Build a word cloud or frequency analysis of the most common terms”
“Make a comparison chart showing engagement differences across content types”
“Generate a correlation heatmap of all numerical features in my dataset”

AI Interpretation Prompts by Level

Beginner:

“What patterns do you notice in this engagement chart? What story does it tell?”
“Looking at this timeline, when are users most and least active?”

Intermediate:

“What cultural or social factors might explain these posting and engagement patterns?”
“How would you design a follow-up study based on these visualization insights?”

Advanced:

“What assumptions about social media behavior are embedded in these visualizations?”
“What are the limitations of this visual analysis approach for cultural research?”

AI Visualization Iteration Prompts:

“What additional visualizations would reveal patterns I might be missing?”
“I notice some outliers in the engagement data. How do I identify and analyze them?”
“How can I create visualizations that tell a cohesive analytical story?”

Session 4: Advanced Analysis and Course Design

Managing Complex Analysis with Jupyter and AI

AI Prompts for Advanced Setup:

“Help me import and set up advanced analysis libraries like numpy, scipy, and scikit-learn”
“What additional libraries would be useful for network analysis or topic modeling?”

Network Analysis of Your Corpus

AI Prompts for Network Analysis:

“Create a function to extract user mentions and build an interaction network”
“How do I analyze the most connected or influential users in my corpus?”
“Build a network graph showing how users interact through mentions and replies”
“Calculate network centrality measures to identify key community members”

Topic Modeling and Advanced Text Analysis

AI Prompts for Topic Analysis:

“Implement basic topic modeling using TF-IDF and clustering on my social media corpus”
“How do I determine the optimal number of topics for my dataset?”
“Create a function that identifies and labels the main themes in my posts”
“How can I validate these computational topic categories against manual analysis?”

Data Quality and Corpus Assessment

AI Prompts for Validation:

“Create a function to assess the quality and representativeness of my social media corpus”
“How do I identify potential biases or gaps in my data collection?”
“What statistical measures help evaluate whether my dataset is sufficient for analysis?”
“Generate a summary report of key characteristics and limitations of my corpus”

AI Iteration Prompts for Advanced Analysis:

“My topic modeling results don’t make sense. How do I debug and improve the parameters?”
“The network analysis shows no connections. Am I extracting mentions correctly?”
“How can I integrate multiple analysis approaches into a coherent research framework?”
“What are the best practices for validating computational results with traditional methods?”

Course Design Framework

Managing Student Jupyter Workflows

Teaching Jupyter Best Practices:

Cell Hygiene: Teach students to organize cells logically
- Setup cells at top
- One function per cell when possible
- Clear outputs before sharing notebooks
AI Iteration Strategy:
- Always test AI suggestions in new cells first
- Keep working versions before experimenting
- Use markdown cells to document AI conversation highlights
Data Management:
- Save processed datasets as CSV files
- Version control for significant changes
- Clear variable names that reflect data content

3-Week Programming Unit Template

Week 1: Text Processing & Pattern Recognition

AI Strategy: Code comprehension and debugging assistance
AI Prompts Focus: “Explain what this code does” and “Fix this error”
Computational Concept: Loops and conditional logic
Humanities Connection: Close reading → algorithmic pattern detection

Week 2: Data Analysis & Visualization

AI Strategy: Conceptual application and method comparison
AI Prompts Focus: “How would I adapt this for my research?” and “What does this pattern mean?”
Computational Concept: Data structures and statistical thinking
Humanities Connection: Quantitative analysis of cultural texts

Week 3: Interpretation & Critique

AI Strategy: Critical evaluation of computational approaches
AI Prompts Focus: “What are the limitations?” and “How does this compare to traditional methods?”
Computational Concept: Algorithm design and limitations
Humanities Connection: Hermeneutics and computational interpretation

AI Integration Best Practices

What Works ✓

Layered AI prompting for different skill levels
Building on prior computational experience
Connecting code generation to analytical thinking
Using AI to focus on concepts over syntax

What to Avoid ✗

AI as a black box students can’t interrogate
Over-reliance without understanding fundamentals
One-size-fits-all AI assistance approaches
Code generation without research context

Disciplinary Applications

Historical Studies

Research Questions:

How do anniversary commemorations reveal evolving historical consciousness?
What patterns in historical discourse show shifting collective memory?

AI Prompts:

“How can computational temporal analysis complement traditional historical methods?”
“What biases might be present in digital historical discourse data?”
“Create code to identify historical references and commemorative language in social media posts”

Literary Studies

Research Questions:

How do narrative conventions appear in social media storytelling?
What forms of community emerge around shared literary texts?

AI Prompts:

“How do we apply close reading techniques to computational analysis results?”
“Generate code to identify narrative structures and literary devices in short-form social media text”
“What’s the relationship between algorithmic pattern detection and literary interpretation?”

Cultural Studies

Research Questions:

How do digital communities negotiate cultural identity?
What role do platform affordances play in cultural practices?

AI Prompts:

“How can computational methods reveal cultural patterns while preserving meaning?”
“Create functions to identify and analyze cultural identity markers in social media text”
“What ethical frameworks should guide digital cultural analysis?”

Quick Implementation Guide

This Week:

Try one AI-assisted coding exercise in your current course
Experiment with different prompt structures for your students
Use AI to help create one programming activity for your discipline

This Month:

Design an AI-integrated assignment using these prompting strategies
Connect with colleagues doing similar experiments
Document successful AI interactions and prompt patterns

This Semester:

Pilot a full unit using these AI assistance strategies
Collect student feedback on AI learning experiences
Share results with DH pedagogy community

AI Prompt Bank for Students

Debugging & Comprehension

“This code isn’t working. Here’s the error message: [paste error]”
“Explain what each line of this code accomplishes”
“How is this different from the approach I used in [previous project]?”
“I don’t understand why this function returns [unexpected result]. Can you walk through it?”

Conceptual Application

“How would I adapt this method for [specific research question]?”
“What other analysis techniques could I apply to this type of data?”
“Connect this computational approach to [disciplinary method or theory]”
“I want to modify this analysis to focus on [specific aspect]. How would I start?”

Critical Analysis

“What are the limitations of this analytical approach for [research context]?”
“What assumptions are embedded in this algorithm or method?”
“How might this computational method bias results toward certain conclusions?”
“What would be a more rigorous approach to validate these findings?”

Code Generation and Development

“Create a function that [specific task] using [specific libraries or approaches]”
“How do I extend this basic analysis to include [additional features]?”
“I need to process [type of data] to extract [specific information]. Help me design the workflow”
“Build a visualization that shows [specific pattern or relationship] in my dataset”

Resources & Follow-Up

Workshop Materials

Starter Jupyter notebook template with AI prompts
Complete AI prompt library organized by skill level
Sample assignment frameworks for different disciplines

Professional Development

Digital Humanities pedagogy networks
AI in education research communities
Programming education best practices

Assessment Strategies

Focus on process documentation and AI interaction reflection
Emphasize interpretation alongside technical execution
Include peer review of AI-assisted work
Measure growth in computational thinking through iterative portfolios