Kail-Fu

Kail-Fu / social-worlds

Public

Social Worlds: Visualizing Social Connections in Captioned Image Corpora

12
1
100% credibility
Found Mar 02, 2026 at 12 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Tools for creating visualizations that reveal semantic similarities and groupings in collections of captioned images from historical datasets.

How It Works

1
📖 Discover Social Worlds

You come across this tool while reading a research paper or watching a video about exploring connections in old captioned drawings.

2
📥 Gather Your Data

Download the spreadsheet of captions and the folder of images to start exploring your collection.

3
🛠️ Set Up Workspace

Follow simple instructions to prepare everything on your computer so it's ready to analyze images.

4
✨ Launch Analysis

Run the quick pipeline that finds similarities between captions and arranges images into meaningful groups.

5
🗺️ Create Visual Maps

Generate scatter plots, tree diagrams, and clusters that show how images relate by their descriptions.

6
🔍 View Results

Open PDF files or upload data to an online viewer to zoom around and see patterns emerge.

🎉 Uncover Connections

You now have beautiful visualizations revealing social worlds and hidden links in your image collection.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 12 to 12 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is social-worlds?

Social-worlds turns Excel sheets of captioned images into visualizations of semantic connections, using the social worlds framework to map arenas of similarity like social worlds arenas theory in organizational contexts. You feed it multilingual captions from datasets such as historical Vietnamese drawings, and it spits out t-SNE/UMAP projections, image-embellished minimum spanning trees via Graphviz, hierarchical dendrograms, radial trees, and PixPlot exports—all via Python CLI commands like sw-similarity and sw-mst. Built on sentence-transformers for embeddings, it solves the pain of manually exploring social links in captioned image corpora.

Why is it gaining traction?

It delivers a ready-to-run pipeline from raw sheets to publication-ready viz, with Makefile targets for full automation and support for English, French, and Vietnamese models. Developers dig the zero-config quickstart for social worlds examples, plus easy coloring by metadata like gender for legitimation processes insights. No alternatives bundle embedding, DR, and graph viz this tightly for social network previews in image sets.

Who should use this?

Digital humanities researchers analyzing social worlds of premodern transactions or children learning to write. Data viz engineers handling OSINT from captioned artifacts. Historians visualizing social democracy in late antiquity via github social media preview tools adapted for corpora.

Verdict

Niche but solid for social worlds arenas map needs—grab it if your captions fit the format, thanks to strong README and video demo. With 12 stars and 1.0% credibility score, it's early-stage research code; fork and extend rather than productionize.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.