NEExT / polblogs /card.md
anomalypoint's picture
Add files using upload-large-folder tool
694fc1f verified
|
Raw
History Blame Contribute Delete
1.49 kB
# Political Blogs (Polblogs)
**Task**: Political-leaning classification
**Size band**: small · **Label type**: semantic
**Label column**: `leaning` · **Converter**: `converters/convert_small_text.py`
Hyperlink network of US political blogs captured before the 2004 election. Label is the blog's political leaning (0 = liberal, 1 = conservative). Community structure correlates strongly with the label.
## Converted graphs (neext/)
| graph | nodes | edges | classes | feature cols | isolated | class counts |
|---|---|---|---|---|---|---|
| default | 1,490 | 16,715 | 2 | 0 | 266 | 0: 758, 1: 732 |
*Conversion notes*: Netzschleuder CSV bundle; 'value' column (0=liberal, 1=conservative) used as label; blog names dropped; directed hyperlinks symmetrized. Contains isolated nodes.
## Source
- [polblogs.csv.zip](https://networks.skewed.de/net/polblogs/files/polblogs.csv.zip) — 93,843 bytes, sha256 `a854433122316724…`, fetched 2026-07-23
**License**: Research use; cite Adamic & Glance
**Citation**: Adamic, Glance. The Political Blogosphere and the 2004 U.S. Election: Divided They Blog. LinkKDD 2005.
- https://networks.skewed.de/net/polblogs
## Caveats
- ~266 isolated nodes; NEExT's filter_largest_component or explicit nodes.csv handling decides their fate.
- Newman's UMich mirror blocks scripted access; Netzschleuder is the reliable source.
---
*Generated by converters/make_cards.py; stats from metadata.json. Raw files: `source/`. NEExT tables: `neext/`.*