Student Coding Questions to an AI Assistant from a 12-Week C Programming Class

A collection of 6,776 questions asked by students to the CodeAid AI assistant during a C programming class from January to April 2023. The dataset was created by majeedkazemi and uploaded to Hugging Face in February 2024. CodeAid, powered by GPT-3, provided scaffolded, teaching-assistant-like responses instead of direct code solutions.

Use Cases

• Analyze student learning patterns and common difficulties based on the natural language questions asked during a programming course.
• Train or fine-tune AI teaching assistants to generate scaffolded responses based on the types of queries students ask.
• Study the interaction dynamics between students and an AI assistant that is restricted from providing direct code solutions.

Strengths

• Contains 6,776 distinct student questions, providing a substantial corpus for analysis.
• Covers a defined 12-week semester, offering a temporal scope for longitudinal study.
• Questions originate from a real educational context where a specific AI tool (CodeAid) was permitted.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known, but the specific content and structure of each question record are unknown.
• Data may reflect temporal and source bias inherent to a single class and AI assistant in 2023.

Provenance

Source

Hugging Face, uploaded by author majeedkazemi.

Collection Method

Questions logged from student interactions with the CodeAid AI assistant during a university C programming class.

Time Range

January to April 2023 (12-week semester).

Freshness

Last updated 2024-02-26 17:01:00; freshness should be verified.

Geography

Region is tagged as 'us'; specific institution is unknown.