Research Study: Stack Overflow Usage Pre vs Post LLM

• Pre LLM time period: December 1st 2018 - September 30th 2020
• Post LLM time period: December 1st 2022 - September 30th 2024
• SQL queries

Data research plan:

Posts:

• Average number of posts
• Average number of posts per day
• Average length of body of post
• Average number of posts per hour of the day

Questions:

• Number of questions posted
• Average score per question
• Average view count per question

Answers:

• Average number of reply's per post
• Number of post reply's
• The average of the number of reply's per day
• Average length of answer
• Average number of votes per top answer

Users:

• Average number of accounts created
• Number of active users
• Average number of dailey users

Quality of responses:

• Record the preferred response
• Record the number of verbs vs adjective used per response.

Popularity of topics discussed

• Determine which topics/coding languages are most searched/tagged in posts.
• Count the number of tags accosiated to each topic.

Important study information to consider:

• A Post is consider to be anything "posted" to Stack Overflow. This includes questions, answers, comments, and forum discussions.
• A Question is consider to be a Post that has no parentId. Therefor is it a question being asked which marks the beginning of a dicussion. It is not in reply to another Post
• An Answer is consider to be a Post that has a valid parentId. Therefor when a user engages in a discussion they are considered to be replying to a Post of type Question.

Installation guide:

1. Install python 3.9
2. Create virtual environment:

   python3 -m venv venv

3. Install requirements

   pip install -r requirements.txt

4. Run the Pre vs Post LLM graph generation seperately
   • Pre LLM:

     python3 -m src.pre_LLM

   • Post LLM:

     python3 -m src.post_LLM

   • Comparison Graphs:

     python3 -m src.comparison_LLM

5. (optional) Run all graph generation simultaniously.

   python3 -m src.run_all

6. (optional) Run pfd generation.

• Pre-LLM:

  python3 -m src.pdf_generation.pre_llm_pdf_generation

• Post-LLM:

  python3 -m src.pdf_generation.post_llm_pdf_generation

Linter

• To run project lint:

  python3 -m black .