diff --git a/src/routes/school/css/+page.svelte b/src/routes/school/css/+page.svelte index b17faef..05263b2 100644 --- a/src/routes/school/css/+page.svelte +++ b/src/routes/school/css/+page.svelte @@ -92,6 +92,70 @@
+ The analysis is based on the United Nations General Debate Corpus + (UNGDC), which contains speeches delivered by countries at the United + Nations over multiple decades. Each observation consists of a speech, a + country, and a year, making it possible to connect political language + with both countries and time. +
++ This makes the dataset especially useful for studying: +
++ In this project, countries are treated as nodes, while mentions from one + country to another form directed edges. A directed edge means that one + country refers to another in a speech. The direction matters, because + diplomatic attention is not necessarily symmetric: a country may talk + about another country much more than it is talked about in return. +
++ Edge color reflects whether the mention is framed in a more positive or + negative way. This allows the network to capture both the volume and the + tone of diplomatic attention. +
+ ++ The graph below shows the full diplomatic mention network built from + the UN speeches. Each node represents a country, and each directed + edge represents one country mentioning another. Edge color reflects + whether the mention is framed in a more positive or more negative + way. Users can adjust the positive and negative sliders to filter + the graph by sentiment strength. Because the full graph contains a + very large number of connections, it provides a useful overview of + how dense and interconnected international political discourse is, + but it is too crowded for detailed country level interpretation. +
+