Pavel.
All projects
Research

ProtoData

Mining and Mapping the Protoweb studies structure and behaviour of the early web using restored sites via a Protoweb proxy. The crawler is class-based (initialization, collection, storage) with optional domain boundaries, then materializes data for SQL and graph analysis.

PythonBeautifulSoupSQLiteNeo4jpandas

Highlights

  • Proxy-aware crawling with domain-boundary and open-boundary modes.
  • Triple storage: raw CSV, relational SQLite (pages/links), Neo4j graph of URLs and edges.
  • SQL/pandas analysis scripts over the structured crawl.
  • Research write-up: “Mining and Mapping the Protoweb.”

Outcomes

Dataset on the order of ~3001 pages across ~40 domains.
Repeatable pipeline from crawl → SQLite → Neo4j for structural questions.

Why it matters

Shows data engineering and systems thinking: careful crawling, storage tradeoffs, and turning messy web archives into analyzable graph and relational models.