Defense
org Protection from attack in military operations. Wikidata
Named in 3290 articles by 136 newsrooms, in 1124 stories, on 14 of the last 14 days, first on 11-07.
Most exclusively tied top 30 · ties that pass the test on the right, most exclusive first
The names that appear with Defense in a large share of their own coverage: this name's own company. Exclusivity runs from 0, two names that meet no more than chance would have them, to 1, two names that only ever appear together. A name that is in everything, a president, is exclusive to no one, so the big names fall away here and the names that belong to this one stay. Only ties that pass the strength test count, so a pair seen twice is not a tie. Places are context and have no page.
| Name | Kind | Exclusivity | Together |
|---|---|---|---|
| Pete Hegseth | person | 0.704 | 2453 |
| Thomas Massie | person | 0.695 | 1005 |
| Department of Defense | org | 0.595 | 834 |
| Defense Ministry | org | 0.546 | 192 |
| Kentucky | gpe | 0.522 | 344 |
| Caribbean Sea | loc | 0.48 | 41 |
| Minab | gpe | 0.477 | 21 |
| Hakeem Jeffries | org | 0.47 | 333 |
| South Korea- | gpe | 0.465 | 100 |
| Koreas | gpe | 0.454 | 64 |
| Venezuela | gpe | 0.45 | 211 |
| Hill | org | 0.434 | 127 |
| Lansing | gpe | 0.424 | 8 |
| Cebu | gpe | 0.418 | 56 |
| World War I | event | 0.401 | 25 |
| Pelosi | event | 0.399 | 15 |
| East Sea | loc | 0.398 | 56 |
| Iran war | event | 0.39 | 304 |
| Capitol Hill | org | 0.388 | 92 |
| Newport | gpe | 0.385 | 11 |
| World War | event | 0.379 | 23 |
| Bundeswehr | gpe | 0.374 | 13 |
| War | gpe | 0.368 | 29 |
| Jeffries | gpe | 0.367 | 84 |
| S. Korea | gpe | 0.365 | 182 |
| Caribbean | loc | 0.364 | 114 |
| Lee Jae Myung | person | 0.363 | 147 |
| US House of Representatives | org | 0.356 | 68 |
| GOP | org | 0.355 | 446 |
| Holocaust | event | 0.355 | 13 |
Most strongly tied the ties chance explains least
How sure the record is that two names appear together for a reason. Strength is a standard statistical test (log-likelihood): above 10.8, chance would produce the pairing less than once in a thousand times. It grows with the number of articles, so the biggest names come first here; the list on the left is where this name's own company shows.
| Name | Strength | Together |
|---|---|---|
| Pete Hegseth | 19773 | 2453 |
| Thomas Massie | 8707 | 1005 |
| Department of Defense | 6004 | 834 |
| Congress | 2527 | 819 |
| Donald Trump | 2429 | 1267 |
| Kentucky | 2366 | 344 |
| Iran | 2021 | 958 |
| Hakeem Jeffries | 1998 | 333 |
| GOP | 1748 | 446 |
| Washington | 1697 | 574 |
| United States | 1572 | 1315 |
| Defense Ministry | 1540 | 192 |
| Iran war | 1435 | 304 |
| Venezuela | 1274 | 211 |
| Senate | 1188 | 520 |
Articles per day
The hubs' circles share of each hub's 30 closest names that are also linked to this one
The eight most-named figures each have an inner circle: the thirty names most tied to them. This is how much of each circle this name is also tied to. A count of shared company in the coverage, nothing more.
Shared with White House: ASEAN, Abdul El-Sayed, Air Force, Alice Huang, Amy Acton, Andy Burnham, Anthony Albanese, Anthropic’s Claude, Armed Forces, Association of Southeast Asian Nations, Barack Obama, Benjamin Netanyahu, Bernie Sanders, Boeing, CIA, COP31, Cabinet, Capitol Hill, Chinese Communist Party, Chuck Schumer.
The year 7079 articles on 139 days, 2025-09-28 to 2026-11-20 · weekly bars · 15 months on record
Left is a year ago, right is this week. Older days come from headlines only, so their counts run lower than live days.
Named together, month by month · by surprise, among that month's articles · names with a page
Stories that name Defense newest edition first · 40 shown
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 20 Nov
- 19 Nov
- 19 Nov
- 19 Nov
- 19 Nov
- 19 Nov
- 19 Nov
- 19 Nov
- 19 Nov
- 19 Nov
- 19 Nov
- 19 Nov
- 19 Nov
- 19 Nov
- 19 Nov
- 19 Nov
Names are found by a language model trained to spot them, joined to one spelling per person, and counted per article. The relations are the model's reading of the sources. Everything else is a count of who was written about together; it says nothing about why.