wikipedia_discussions_de — German Wikipedia discussion replies (2001-2026)

Description

Directed networks of replies between the users of the German Wikipedia in the discussion threads of the talk pages of its articles, extracted from two dumps of the current versions of all pages: 2015, from the dump of 11 February 2015, and 2026, from the dump ("MediaWiki Content File Export") of 1 October 2026, which reflects the state of the talk pages at that time. Each node is a registered user, and each directed edge from user A to user B is a comment of A in reply to a comment of B, so that users who replied to each other several times are connected by multiple edges; users may reply to themselves.

Only the talk pages of articles (namespace "Diskussion", including their archive subpages) are considered, not those of users, projects, templates, etc. The replies are inferred from the wikitext of the current version of each page with the method of N. Geilen and J. Kunegis (University of Koblenz–Landau), reimplemented here: each section heading starts a new thread; within a thread, a comment is a line that starts with colons (its indentation level is the number of colons), that starts with "--" (taking the indentation level of the previous line), or that contains a link to a user page ([[Benutzer:...]], [[Benutzerin:...]] or [[User:...]]). The author of a comment is the user of the last such link of its line (the signature), and its date the last date in the format of the signatures of the German Wikipedia ("12:34, 5. Okt. 2010") of that line; a comment without a date takes that of the closest preceding comment of lower indentation that has one. Comments without an author (including those signed by unregistered users, identified by their IP addresses, and those marked as unsigned with the templates {{unsigned}} and {{unsigniert}}), without a date, or dated before 15 January 2001 or after the date of the dump are left out. Each remaining comment of a thread, except the first, is a reply to the closest preceding remaining comment of the thread with a lower indentation level, or with indentation level zero. Since only the current versions of the pages are used, discussions removed from the talk pages (and not archived) are not included, and since comments are recognized from their signatures, some replies are missed or attributed approximately.

User names are normalized as MediaWiki page titles (underscores replaced by spaces, first letter capitalized), so that different spellings of the link to the same user page give the same node.

The networks have the following properties:

(Node) name: the user name.
(Edge) time: the time of the reply, from its signature (Unix time; the signatures give it to the minute, in German time, which is converted to UTC).
(Edge) inferred_time: whether the comment had no date of its own, so that its time is that of the closest preceding comment of lower indentation.
(Edge) page: the title of the talk page.
(Edge) thread: the title of the section (empty for comments before the first section heading of the page).

Tags

Social Online Communication Unweighted Multigraph Timestamps Metadata

Added

2026-10-27

Citation

"Wikimedia Foundation. Wikimedia downloads." dumps.wikimedia.org

Upstream URL OK

https://archive.org/details/dewiki-20150211
https://dumps.wikimedia.org/other/mediawiki_content_current/
https://github.com/kunegis/konect-extr/tree/master/extr/wikipedia-discussions

Upstream license

Creative Commons Attribution-ShareAlike 4.0 (CC BY-SA 4.0) and GFDL, https://dumps.wikimedia.org/legal.html

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
2015	89,893	2,445,894	27.21	348.61	303.73	32.62	0.19	0.24	13	0.98	Directed	Unipartite	name	time inferred_time page thread	30.6 MiB	42.6 MiB	40.9 MiB	47.6 MiB
2026	126,559	3,987,918	31.51	507.88	408.06	106.32	0.21	0.30	13	0.98	Directed	Unipartite	name	time inferred_time page thread	52.1 MiB	70.9 MiB	68.2 MiB	78.6 MiB

Ridiculograms^*

^* These are automatically generated force-directed visualizations, and can be quite meaningless for networks both big and small. They should not be taken seriously as sources of scientific insight. See here for a discussion.