---
title: "Research Papers public analytics snapshot"
description: "Statistics, provenance, and downloadable JSON for the public research snapshot."
canonical: "https://papers.highsignal.app/data"
last_updated: "2026-09-01"
---

# Research Papers public analytics snapshot

The public snapshot describes 478,991 papers,
8,981,698 citation edges, and the static analytics
exports generated from the operator-side ClickHouse corpus.

## Sources

- arxiv: 399,902 papers
- biorxiv: 63,752 papers
- openreview: 30,327 papers
- medrxiv: 17,332 papers

## Downloadable JSON

- [summary.json](https://papers.highsignal.app/data/summary.json): Corpus ingest summary: paper counts, edge counts, storage size
- [top_papers.json](https://papers.highsignal.app/data/top_papers.json): Top 200 papers by PageRank and Katz centrality
- [top_cited_works.json](https://papers.highsignal.app/data/top_cited_works.json): Most-cited works across the corpus
- [hot.json](https://papers.highsignal.app/data/hot.json): 100 hot papers: citation velocity, review rating, and PageRank
- [sleepers.json](https://papers.highsignal.app/data/sleepers.json): 100 reviewer-loved early papers with low citation counts
- [communities.json](https://papers.highsignal.app/data/communities.json): 64 semantic communities with anchor papers and labels
- [embedding_clusters.json](https://papers.highsignal.app/data/embedding_clusters.json): Embedding cluster assignments from MiniLM
- [temporal.json](https://papers.highsignal.app/data/temporal.json): Papers per year, cites per year, and top paper per year
- [tag_rating.json](https://papers.highsignal.app/data/tag_rating.json): Mean reviewer rating by extracted topic
- [tag_cooccurrence.json](https://papers.highsignal.app/data/tag_cooccurrence.json): Topic tag co-occurrence matrix
- [review_venues.json](https://papers.highsignal.app/data/review_venues.json): Peer-review venue summaries and review counts
- [review_rating_distribution.json](https://papers.highsignal.app/data/review_rating_distribution.json): Rating distribution per review venue
- [ch_sources_summary.json](https://papers.highsignal.app/data/ch_sources_summary.json): Paper counts by source: arXiv, OpenReview, bioRxiv, and medRxiv
- [top_authors.json](https://papers.highsignal.app/data/top_authors.json): Most prolific authors in the corpus
- [top_hosts.json](https://papers.highsignal.app/data/top_hosts.json): Top hosting domains for paper URLs
- [cycles.json](https://papers.highsignal.app/data/cycles.json): Citation cycle detection results
- [abstract_clusters.json](https://papers.highsignal.app/data/abstract_clusters.json): Abstract-level clustering results
- [host_categories.json](https://papers.highsignal.app/data/host_categories.json): Host URL categorization

## Provenance

The corpus combines arXiv, OpenReview, bioRxiv, and medRxiv metadata. Citation
analytics, semantic communities, reviewer signals, and static exports are built
from the repository's documented ingestion and ClickHouse pipelines. JSON
exports are data resources and are intentionally separate from the HTML sitemap.

## Public-site analytics

PostHog measures bounded product events and Microsoft Clarity measures aggregate
usage and session replay on this public, no-account research library. The
semantic-search and cited-answer sections are explicitly masked, so their query
text and rendered results are not uploaded to Clarity.