Work / Supporting engineering
Documented implementationSocial Trend Analysis Workflow
From collected posts to grouped topics and summaries.
A Python proof of concept that separates social data collection, text preparation, embeddings, clustering, and topic summaries into an inspectable workflow.
From posts to topics
I organized a Threads analysis POC into collection, text cleanup, embedding generation, K-means grouping, and summary generation. The public repository shows how those stages connect, making it possible to inspect how collected text becomes a topic report.
Following a topic back to its inputs
The useful review path runs backward from a summary to its grouped text and underlying posts. Looking at each stage helps distinguish a collection issue from a grouping or summarization issue.
Current status and limits
This is a documented POC, and current platform collection behavior has not been revalidated. A homepage sample represents that collection session, not the whole platform. The code also supplies demonstration fallback data when embedding requests fail, so a report needs its execution context before it can be interpreted as live analysis.