Source Profile — GDELT TV Explorer
Information Sources

Source Profile — GDELT TV Explorer

Skip to main content
< All Topics
Print

Source Profile — GDELT TV Explorer

Posted
Updated
ByPatriot University
0 out of 5 stars
5 Stars 0%
4 Stars 0%
3 Stars 0%
2 Stars 0%
1 Stars 0%

What it is

GDELT (Global Database of Events, Language, and Tone) TV Explorer is a research-grade interface to the Internet Archive’s TV News Archive, providing programmatic search and analysis of TV closed-captioning text from 2009 onward.

URL: https://api.gdeltproject.org/api/v2/tv/tv (API); https://blog.gdeltproject.org/announcing-the-gdelt-television-news-tv-explorer/ (overview)

Coverage

  • All US national news networks captured by Internet Archive TV News Archive
  • 2009-present, continuous since 2010
  • Closed-captioning text indexed at 15-second granularity
  • Time-coded video links back to Internet Archive

Strengths

  • Programmatic API access: REST API with JSON responses
  • 15-second segment granularity: Locate specific moments in broadcasts
  • Topic and frequency analysis: Built-in tools for “how often was X mentioned on which networks”
  • Timeline visualization: API returns time-series data for trend analysis
  • Free for research use: No API key required for basic use
  • Cross-network analysis: Single query can compare CNN, MSNBC, Fox News mentions

Limitations

  • Closed-captioning quality: Inherits all auto-transcription errors from IA TV News Archive
  • Not full-transcripts: Returns segments around hits, not full broadcasts
  • Coverage of Trump speech vs Trump coverage: API doesn’t distinguish — additional filtering required to isolate Trump’s own speech
  • Rate limits: Public tier has request rate limits; bulk research may require coordination
  • No speaker identification: Captioning doesn’t always identify speakers
  • API can be unstable: Occasional outages and rate-limit changes; verify current status

API Endpoints

The TV Explorer API supports several modes:

Mode Use
mode=clipgallery Returns clips matching a query with thumbnails and metadata
mode=timelinevol Time-series of mention volume
mode=timelinevolnorm Time-series normalized by total airtime
mode=stationchart Per-network breakdown of mention volume
mode=timelineshare Share-of-coverage time series
mode=clip Returns full clip detail with caption text

Ingestion Approach

For the speech corpus pipeline, GDELT TV Explorer is most useful for:

  1. Verifying broadcast coverage of known events:
  • Query: “Trump rally [date]”
  • Use to confirm event happened and was televised
  1. Identifying networks-by-event coverage patterns:
  • For each major Trump event, which networks ran the speech live? Which ran clips?
  1. Topic-frequency tracking:
  • Time-series of “Trump” mentions by network
  • Spike detection for newsworthy events
  1. Cross-reference for primary corpus:
  • For any speech in primary corpus, confirm at least one network’s broadcast captured the same content

For ingesting Trump-speech segments specifically:

  1. Construct queries: "trump" near "speaking" or "trump rally" or "trump news conference"
  2. Iterate by date range: Daily or weekly windows
  3. Per-item fields:
  • source_id: GDELT clip ID + IA TV News Archive ID
  • title: Show name + segment timestamp
  • body: Closed-captioning text for the segment
  • published_at: Broadcast date + time
  • url: GDELT permalink + IA video URL
  • doc_type: tv_segment
  • extra_metadata: Network, show, segment timestamps, closed-captioning quality flag
  1. Mark transcript-quality: All as auto_generated_captions; do not use for word-level lexical analysis without verification
  2. Rate limit: 1-2 req/sec recommended; respect rate-limit response codes

Use Cases for Speech Corpus

  • Coverage volume tracking: “How often did Fox News quote Trump on immigration in 2025?”
  • Cross-network comparison: “How did MSNBC, CNN, and Fox News cover the [specific event]?”
  • Event-existence verification: Cross-check that a claimed event was actually televised
  • Frame analysis input: TV news show coverage of Trump speech is itself a research subject

Citation Convention

GDELT TV Explorer, “[Query],” accessed [date]. https://api.gdeltproject.org/api/v2/tv/tv?query=[encoded_query]. Internet Archive TV News Archive segment: https://archive.org/details/[id]#start=[timestamp].

API Best Practices

  • Always check API status before bulk operations
  • Use timeline-mode API rather than per-clip API for coverage-volume questions (more efficient)
  • Cache aggressively — query results are stable
  • Document API version at time of query (GDELT API has had v1 and v2; behavior differs)

Maintenance

  • Refresh: Daily query for recent events
  • Backfill: One-time historical pull for major events 2015-present
  • API-version checks: Quarterly verification that endpoints and parameters haven’t changed

See Also

  • sc-overview.md
  • sc-source-internet-archive.md (the underlying TV News Archive)
  • sc-source-american-presidency-project.md
  • sc-source-rev-transcript-library.md
Was this article helpful?
0 out of 5 stars
5 Stars 0%
4 Stars 0%
3 Stars 0%
2 Stars 0%
1 Stars 0%
5
Please Share Your Feedback
How Can We Improve This Article?
Table of Contents

Related Articles