Are there geographical spheres of interest?
Let us consider all geographical entities used as meta-keywords. For both outlets,
between 40 to 50% of news content tagged with geo-related keywords refer to the
newspaper's own country. The rest of this coverage is split across other countries,
as shown on the left. The frequency at which certain geo-political entities are
referred to translates into relevance for the reader. This relevance is constructed
by following certain patterns. For example, these keywords are set at different
granularities. At a low level we have names of cities, like Berlin or New York,
sometimes even names of neighborhoods, like Manhattan (NY). At a higher level, we
have names of states (Baden-Wuttemberg, Ohio), countries (Bolivia, Italien,
Greenland), and supra-national geo-political or geographical entities (Red Sea,
Atlantic Ocean). If we reduce the granularity by aggregating all localities with
their country and assign each region and country to macro-regions, we can observe
how newspapers have different areas of interest.
The resulting abstraction is shown in the visualization on the right, with the world “split” into areas of interest based on how much either Die Zeit or The New York Times cover a specific country.
How to read the chart:
Notes:
- 1. The count of keywords related to the newspapers’ respective countries is excluded.
- 2. Keywords grouped in the “Other” category refer to geographical landmarks or historical geopolitical entities