Overview of Modules

scraper.py

• TokenScraper

  • Fetches token metadata (creator, mint time, etc.) via Solscan API.
  • Falls back to a local step_0.json file if API calls fail.

• MinterTransferScraper

  • Performs a breadth-first search (BFS) from the token’s minter address.
  • At each “step,” retrieves outgoing transfers, forms child nodes, and saves progress to JSON (e.g., step_1.json).
  • Supports a configurable maximum depth (max_steps) and brief delay between API calls.

utils.py

• serialize_graph / deserialize_graph
  • Converts a tree of AccountNode objects into JSON-friendly structures (and back again).
• save_graph_state_to_json
  • Saves the BFS “root” and current “frontier” (list of addresses) in a single JSON file for incremental scraping or later reloading.

---

Key Conclusions

1. Meme Tokens = Star-Like Flows

   • Large address sets and negative out→in degree correlations indicate a few “whales” distributing tokens widely.
   • One-directional edges dominate, leading to many small strongly connected components.

2. Fundamental Tokens = Smaller BFS Footprints

   • Minimal direct user-to-user transactions uncovered via BFS.
   • Usage likely happens off-chain, via custodial services, or through specialized smart contracts.

3. Network Metrics Reveal Hype vs. Utility

   • Low clustering and broadcast-style flows are typical of hype-driven assets.
   • Steady, functional tokens might show sparse on-chain distribution if most operations occur elsewhere.

Overall, systematic graph-building from the minter address highlights how token type (meme vs. fundamental) can profoundly shape on-chain footprints, even on a high-throughput blockchain like Solana.