This is a prototype. The analysis may be distorted by errors in the code, and users should independently verify any findings.
This dashboard draws on every article published in The New York Times from January 2000 to the present, retrieved through the NYT Archive API. Data is updated monthly.
Latest update: With the help of users who identified duplicate data in the API, a process has been added to remove duplicate articles and subdivide blog and non-blog content, which resulted in some slight changes in the total material displayed.
The API includes both standard articles and blog content — the latter hosted at *.blogs.nytimes.com subdomains and dealbook.nytimes.com. Blog posts were typically much shorter (averaging 100–460 words vs. 600–900 for non-blog articles), published in higher volume, often without an author name attached, and probably did not reflect newsroom hiring or traditional story assignments. DealBook began appearing in the API in 2006, and the collection of blogs (ArtsBeat, CityRoom, The Caucus, The Lede, Economix, Bits, Wheels, Fifth Down, Straight Sets, and more) peaked at nearly 56% of all indexed articles in 2009. The spike in the second half of 2009 and its abrupt end in November reflect a labeling change in the API — posts that had been filed under their parent sections (Sports, Arts, etc.) were temporarily assigned a "Blogs" section tag, inflating counts during that window. Blog indexing then declined steadily as the NYT wound down its blog properties between 2013 and 2018; by 2019, essentially no content is published at blog subdomains. As that wind-down progressed, writers moved to the main site and their articles shifted to regular nytimes.com URLs — appearing as a rise in non-blog counts between roughly 2013 and 2015 even as total output held steady; total article volume, not just the blog share, begins its genuine decline after blogs fully phase out around 2017.
Article counts in this dashboard are also affected by a URL-duplication problem in the API. Beginning around May 2006, the NYT introduced a new slug-based URL format alongside its old date-coded URLs, causing the API to index many articles twice — once under each URL. A similar issue affects blog content in 2009: blog posts and DealBook articles cross-posted to the main site were often indexed under two or three separate URLs simultaneously. To correct for this, this dashboard removes articles that share the same headline, publication date, and a closely matching word count (within 10%), retaining the shortest URL in each duplicate group. Approximately 62,000 duplicate entries were removed through this process, concentrated in 2006 and the 2008–2009 period of peak blog activity. A separate discontinuity is visible in January 2007, when non-blog counts drop by roughly 2,000 articles per month; analysis of the affected content suggests the API stopped consistently indexing short-form items — particularly letters to the editor and Metro section briefs — rather than any reduction in publishing. The median article word count in the data jumps from about 550 to 660 words at this breakpoint, consistent with short items being dropped from the index.
Each author is assigned a primary section based on where they have the most published articles, used to filter and sort the Reporters table. Author profiles also list all sections in which the author has published, reflecting the full range of their work. Primary section is a rough proxy for desk assignment and may not be accurate for authors who write regularly across multiple sections.
The Reporters and Sections tables include a toggle to "limit to inferred reporters," which filters out bylines that appear to belong to photographers, videographers, podcast producers, illustrators, and similar contributors rather than traditional text reporters. The inference is based on combinations of shared-byline rate, average word count, and primary section (e.g., someone with 90%+ shared bylines, near-zero word counts, and a primary section of "Multimedia/Photos" is unlikely to be a text reporter). The heuristic is imperfect — some edge cases will be miscategorized in either direction — and is intended only as an exploratory filter, not a definitive classification.
The topics shown on author profiles are derived from the NYT's own subject tags on each article. Topics are scored by how disproportionately an author covers them relative to the full corpus — a topic must appear at least twice as often in an author's work as the corpus-wide average to qualify. Authors with few articles or articles lacking subject tags may have no beats listed. Subject tags are assigned by the NYT, not by this dashboard.
Author name variants (e.g., "Jonah E. Bromwich" and "Jonah Bromwich") are merged when they share first and last names with compatible middle names. Cases where the same reporter appears under both a shorter and fuller byline form — each with many articles — are resolved manually after cross-checking beats, section, and year range. Institutional bylines (The Associated Press, The Editorial Board, Reuters, etc.) are excluded from author rankings.
Geographic data for 2001–2003 is unreliable. In the aftermath of September 11, the NYT's metadata system appears to have applied a Washington, D.C. geographic tag as a proxy for national-significance stories — stories that would not normally carry a geographic tag at all. D.C.'s share of World and U.S. section articles spiked from roughly 0.1% in 2000 to over 6% in 2001, then gradually returned to normal by 2003. The original metadata has been retained rather than corrected, as any automated fix would be speculative.
The States tab draws only from articles in the "U.S." section (the national desk). The World tab draws only from articles in the "World" section (the international desk). Coverage of these geographies that appears in other sections (Opinion, Business, etc.) is not included. Both tabs use the API's geographic keyword field; city-level tags (e.g., "Chicago (Ill)") are mapped to their parent states. Per-capita figures use 2025 population estimates.
New York and New Jersey coverage primarily appears in the dedicated "New York" section rather than the national "U.S." section. The States tab defaults to U.S.-section data only; a checkbox at the top of the tab allows adding New York section coverage for NY and NJ (counting only articles with explicit geographic tags — untagged "New York" section articles are excluded).
The NYT has renamed, merged, and discontinued sections over the years. To preserve thematic continuity, this dashboard applies the following merges:
Technical categories, product sections, branded-content labels, and small stubs (newsletter aggregators, discontinued apps, TV tie-ins, blogs without a clear section home) are excluded entirely from the Sections tab. This includes the "Archives" section (predominantly paid death notices, not reported content), "Booming" (a baby-boomer lifestyle blog, 2012–2014), "UrbanEye" (a NYC events newsletter), "Guide" (a going-out listings service), and newsletter and podcast aggregators.
Sports coverage declined sharply after the NYT acquired The Athletic in 2022; Athletic content is not included in the Archive API.
Persons and organizations drawn from NYT keyword metadata. Check names to compare coverage over time; click a name to see details.
Loading subjects data…
Search NYT subject tags to find which journalists cover a topic. Tags are drawn from the NYT's own metadata.
Includes major editorial sections. A few minor or discontinued categories are excluded from display. Where the NYT renamed a section while preserving its editorial identity, data from the old and new names are merged — for example, "Fashion & Style" and "Fashion" are counted together under "Style." See the About tab for further details.
Draws on geographic tags from NYT metadata exclusively in articles from the "World" section. Tagging coverage appears to have been less consistent in 2000–2003.
Draws on geographic tags from NYT metadata in the "U.S." section and, optionally, "New York" section to encompass local coverage.
| to search for multiple terms in one line (e.g. iraq | iran)