GDELT Media Agenda Lab

Attention Timeline

Share of monitored online coverage matching the query. Live DOC data reports normalized coverage share when available.

Demo data

Where The Agenda Is Loudest

In live mode this uses source-country or language fields when GDELT returns them; otherwise it shows the demo lens.

Evidence Cards

Recent or representative stories used to interpret spikes.

Television Query Setup

GDELT TV searches station transcript clips. Normalized percentages are best for comparing outlets with different monitored airtime.

Demo data

Stations

TV date anchor

Normalized Airtime Timeline

Percent of monitored clips on each station matching the query.

Top Shows

Programs where the query appears most often.

TV Evidence

Demo examples explain how to interpret clip-gallery results when a live response is available.

Why This Makes A Good Classroom Lab

Students can move from a raw keyword to a defensible claim about media attention.

1. Attention is normalized, not just counted

DOC timeline volume reports coverage as a percent of the global monitored news stream, which makes time comparisons less naive than raw article counts.

2. Tone should be treated as a distribution

The GDELT docs emphasize tone histograms because a single average can hide whether coverage is actually clustered, neutral, or polarized.

3. TV comparisons need airtime normalization

The TV API recommends normalized clip percentages for comparing stations and periods, since monitored airtime and programming mix differ across stations.

API Query Pattern

The page builds URLs from the current query and lens, then queues requests at least five seconds apart.

Interpretation Caveat Media attention is not public opinion

GDELT measures what monitored media says, where, when, and with what machine-coded tone. It should be interpreted as a media agenda signal, not a survey of what people believe.