Dataset Card generated by Kimi K2.6

Dataset Card for tildes_260503

Dataset Summary

A structured crawl of forum topics and comments from Tildes, a non-profit community site. The dataset contains ~68,000 topics and ~666,000 comments with full thread structure (replies, nesting, timestamps, votes).

Tildes is organised into groups (e.g. ~tech, ~games, ~music) and uses a tree-structured comment system with visible vote counts and optional topic tags.

---

Dataset Details

Dataset Sources

• Source Website: https://tildes.net

---

Uses

Direct Use

Suitable for:

• Academic research into forum dynamics and community moderation
• Linguistic analysis of informal web discourse
• Studying thread/reply structures and voting patterns in modern web forums
• Analysis of topic-tagging behaviour

---

Dataset Structure

Schema

Column	Type	Description
topic_id36	string	Unique topic identifier (base-36)
group	string	Tildes group (e.g. ~tech, ~games)
title	string	Topic title
author	string	Pseudonymous username
posted_at	string	Topic timestamp (ISO-8601)
votes	int64	Vote count shown on the topic
link_url	string or null	External link URL (null for text-only topics)
tags	list	Topic tags
body	string	Topic body text (empty for link-only topics)
posts	list	Array of all comments in the topic

Posts Struct

Column	Type	Description
post_id	string	Unique comment identifier
position	int64	1-based position in flat traversal order
depth	int64	Nesting depth (0 = top-level reply)
parent_post_id	string or null	ID of the parent comment this replies to
author	string	Pseudonymous username
posted_at	string	Comment timestamp (ISO-8601)
votes	int64	Vote count shown on the comment
body	string	Comment body text

Statistics

Metric	Value
Topics	68,122
Comments	665,705
Date range	2018-04-26 to 2026-05-03
Mean comments per topic	9.8
Median comments per topic	5
Longest topic	730 comments
Max nesting depth	38
Groups represented	49
Topics with tags	67,534 (99.1%)
Link-only topics	51,960 (76.3%)
Unique topic authors	3,634
Unique comment authors	10,446
Blank comment bodies	2,300 (0.3%)
Mean comment body length (non-blank)	538 characters
Median comment body length (non-blank)	316 characters
Mean topic body length (non-blank)	942 characters
Median topic body length (non-blank)	466 characters

The dataset is heavily skewed toward a few large groups: ~tech accounts for 8,915 topics (13.1%), followed by ~games (6,358) and ~music (6,343).

---

Dataset Creation

Curation Rationale

Tildes is a moderately trafficked discussion platform with a strong emphasis on quality conversation and community-driven moderation. It serves as a case study in a modern, non-corporate alternative to mainstream social media, with visible vote counts, hierarchical threading, and optional topic tags.

Source Data

Data Collection and Processing

• Tool: Scrapy
• Scope: All topics discoverable via the public RSS feed (topics.rss)
• No filtering was applied to content; all crawled topics and comments are included

Who are the source data producers?

Tildes users posting under pseudonymous usernames.

---

Bias, Risks, and Limitations

Content Risks

• Potentially offensive, inflammatory, or politically charged language
• Occasional misinformation presented as fact
• Discussions of sensitive personal, medical, or political topics

Structural Limitations

• The crawl was conducted on 2026-05-03 and includes no content posted beyond that date
• Vote counts are snapshot values at crawl time and may have changed since

Personal and Sensitive Information

• No systematic PII detection or redaction was performed
• Usernames are pseudonymous but may occasionally contain identifying information
• Posts may contain references to real individuals, institutions, or events
• The dataset should be treated as containing potentially sensitive information

Licence and Redistribution

No licence was identified for the source forum content. This dataset is therefore marked as license: unknown.

The dataset consists of scraped public forum posts authored by third parties. The dataset publisher does not claim ownership of the underlying post text and does not grant additional rights beyond any rights users may already have under applicable law.

Users are responsible for determining whether their intended use is lawful, including with respect to copyright, privacy, data protection, and platform terms.