Skip to content

readme.md

Latest commit

 

History

History
executable file
·
19 lines (15 loc) · 785 Bytes

readme.md

File metadata and controls

executable file
·
19 lines (15 loc) · 785 Bytes

Topic & Emotion Analysis

data prepare

Ensure the UGC JSON file named as ugc_all_instance.json, the data form should be:

{
  "instance1.social": ["post1", "post2", "post3", ...],
  "instance2.com": ["post1", "post2", "post3", ...],
  ...
}

run prepare_batch.py

Generate classification prompts and requests. This will generate request files for each instance inside ./dataset/batch_files/*_batch_requests.jsonl.

Submit the *_batch_requests.jsonl files to OpenAI’s Batch API. Once completed, save the results to: ./dataset/batch_output/*_output.jsonl.

run visualize.py

Analyze and visualize the topic distribution and emotional sentiment of posts from two instances (mastodon.social and chaos.social) using pie charts and a comparative bar chart.