How Milky Way reads a country

1. Two samples. From inside: articles by media headquartered in the country, about the country. From outside: articles about the country by media headquartered anywhere else. All open-access, collected .
2. One list of topics. Every article is filed under one topic, and both samples use the same list, so they can be compared.
3. The lens gap — not a score. For each country, a few cards name what one press covers that the other leaves out. Each card gives the subject, one factual sentence, the number of articles on each side and links to them. Select a card to see its articles.

   Silence abroad

   A subject present in the domestic press and absent from the foreign press sampled here.

   Blind spot at home

   A subject present in the foreign press and absent from the domestic press sampled here.

   Same story, different angle

   A subject covered on both sides, where each side reports different aspects. The card says what each side reports, attributed to the outlets.

4. How gaps are found.
5. Other languages. Headlines not written in English are machine-translated into English by DeepL; the original headline is shown in small type underneath.
6. Neutral by construction. Outlets are labelled by ownership — public, private, agency, state-owned or state-funded — never by political leaning. Summaries attribute claims to the outlet that made them, and every summary links to its sources.

Limits, stated plainly

Samples are small — articles a side — and drawn only from open-access sources: outlets that refuse automated access are missing. “Absent” means absent from this sample, not from a country’s whole press. Matching by words can miss a subject written differently, and standfirsts are not translated. Machine translation can be imperfect: the original headline is always shown. The edition is refreshed automatically every few hours.