Polis tends to split into two diagrams when there are more votes This is because a linear PCA will not be able to clearly discriminate high-dimensional distributions no matter how hard you try. Then take the data in that state and put it into UMAP in Talk to the City. What PCA cannot separate, UMAP can. Just do TTTC-like cluster description generation on the separated ones.
I thought of the Map metaphor when the discussion on Polis came up.
Revisit after the election as I don’t feel like making it now.
- Visualizing Ideological Vectors
- You could try this.
Another data source - Opinion cluster analysis of 110,000 people
This page is auto-translated from /nishio/Polis+TTTC的広域マップ using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.