Who leads the debate on waste, and what is it about?
The Danish waste sector is debated by many different actors. We have gathered 80,401 public documents into one map. They include news articles, opinion pieces, social media posts and much more. The map shows which topics take up space and who takes part in the debate.
One point per contribution to the debate
The map holds 80,401 documents, among them news articles, opinion pieces, press releases, consultation responses, parliamentary questions and social media posts. Each document is placed according to what it is about, so documents on related topics sit close together.
The documents fall into 254 named clusters. Points can be coloured by topic, document type or sender type, and a time filter at the bottom of the map shows what was written in a given period.
Where the documents come from
All of the material is publicly available. Most of it is how the media cover waste, from national newspapers to trade and local outlets, along with the opinion pieces they carry. The rest is press releases and reports from companies, municipalities and organisations, questions and cases from the Danish Parliament, items from municipal agendas and public social media posts.
Altogether the documents come from more than 2,000 sources. The oldest date from 1986, but most are from recent years, when sorting and producer responsibility have dominated.
The actors behind the debate
A register of 2,058 actors links the documents to the politicians, municipalities and waste companies, private businesses, interest groups, researchers and media who speak or are mentioned.
The actor map places the actors according to what they talk about. The network shows who appears in the same documents, revealing both the camps in the debate and the actors who connect them.
How the map is made
A language model that understands Danish reads the title and summary of each document and turns the content into numbers. Two documents on the same topic get similar numbers, even when they do not use the same words.
The numbers are then spread across the map, so the distance between two points matches how closely related the content is. The groups that form on their own are the clusters, and we have read through them and given them names.
We use UMAP to spread out the numbers, HDBSCAN to find the clusters and the Alexandra Institute's Danish sentiment model to assess tone.
The map keeps growing
New documents are collected and added to the existing map, so the points stay where they are and new topics get their own clusters. The map was last updated on 17 October 2026.
If something is missing, or an actor is described incorrectly, please let us know at shn@datara.dk.