hacker_news — Hacker News discussions (2006-2026)

Description

Networks of the discussions on Hacker News, the news aggregation and discussion website of the startup accelerator Y Combinator, from its first item in October 2006 to the end of July 2026, from the complete archive of the items of the site (stories, comments, polls and jobs) published on Hugging Face by Open Index, a mirror of the official Hacker News API (via the ClickHouse public dataset). There are two networks:

replies: each node is a user, and a directed edge from user $i$ to user $j$ is a comment of $i$ replying to an item (a story, a comment, a poll or a job posting) of $j$, with one edge per comment (a multigraph). Users replying to themselves give self-loops.
discussions: a bipartite network of users and stories (including polls and job postings), where an edge connects a user to a story if the user commented somewhere in the discussion of the story (the tree of the comments below it), weighted by the number of comments. Only the users and stories with such comments are included.

Items deleted by their authors or by the moderators have no author in the data: the comments that are deleted, or that reply to a deleted item, are left out of replies, and the deleted comments are left out of discussions (the comments in a discussion below a deleted comment are kept). Comments killed by the moderators or flagged by users ("dead") are included and marked. Comments whose parent item is not in the archive are left out. The texts of the items, the URLs of the stories and the poll options are not included. The archive is updated continuously; the data are those of a pinned revision (of 23 August 2026), and the last, incomplete month (August 2026) is left out.

The networks have the following properties:

(Node) name: the user name, or, in discussions, the title of the story (for stories).
(Node) is_user (discussions only): whether the node is a user (otherwise a story).
(Node) item (discussions only): the item number of the story (-1 for users), as in its address https://news.ycombinator.com/item?id=<item>.
(Node) author (discussions only): the user name of the author of the story (empty for users).
(Node) kind (discussions only): story, poll or job (empty for users).
(Node) score (discussions only): the score (points) of the story (0 for users).
(Node) time (discussions only): the submission time of the story, in Unix time (-1 for users).
(Edge) time: in replies, the time of the comment; in discussions, the time of the first comment of the user in the discussion; in Unix time.
(Edge) comment (replies only): the item number of the comment.
(Edge) parent (replies only): the item number of the item replied to.
(Edge) story (replies only): the item number of the story at the root of the discussion.
(Edge) dead (replies only): whether the comment is dead (killed by the moderators or flagged).
(Edge) weight (discussions only): the number of comments of the user in the discussion.

Tags

Social Online Communication Multigraph Timestamps Temporal Bipartite Weighted Metadata

Added

2026-10-28

Citation

Y Combinator, "Hacker News API." github.com
Open Index, "Hacker News - Complete Archive." Hugging Face dataset (2026) huggingface.co

Upstream URL OK

https://news.ycombinator.com/
https://huggingface.co/datasets/open-index/hacker-news

Upstream license

ODC-BY, https://opendatacommons.org/licenses/by/1-0/

Networks

Tip: click on a column header to sort the list, or hover over it for a description.

Column legend

$\left<k\right>$: Mean degree
$\sigma_k$: Standard deviation of the degrees
$\lambda_h$: Largest eigenvalue of the Hashimoto (non-backtracking) matrix
$\tau$: Random walk mixing time
$r$: Degree assortativity
$c$: Global clustering coefficient
$\oslash$: (Pseudo-) diameter
$S$: Relative size of the largest component
NPs: Node properties
EPs: Edge properties

Name	Nodes	Edges	$\left<k\right>$	$\sigma_k$	$\lambda_h$	$\tau$	$r$	$c$	$\oslash$	$S$	Kind	Mode	NPs	EPs	gt	GraphML	GML	csv
replies	1,136,463	41,545,062	36.56	724.43	2240.77	539.09	0.01	0.18	14	0.72	Directed	Unipartite	name	time comment parent story dead	578.1 MiB	657.6 MiB	627.7 MiB	744.0 MiB
discussions	2,581,829	27,344,811	21.18	140.63	458.21	1069.22	-0.11	0.00	20	0.93	Undirected	Bipartite	name is_user item author kind score time	weight time	209.6 MiB	269.3 MiB	270.2 MiB	315.2 MiB