prev Weekly journal 2024-10-28~2024-11-06 What did you do last week?
- Writing a summary of [in-house TTTC - Examples of Broad Listening Applications within Cybozu
- Remaining work on the public opinion map
- I was wondering if it would be possible to create a detailed map for the ones with the most people.
- At first I was thinking of keeping PCA and increasing the k of k-means, but then I decided to try [UMAP the opinion vectors
- Past Case Studies - Opinion cluster analysis of 110,000 people did an analysis fixing K=5, not very interesting - Polis-like visualization of the 2022 House of Councillors election would be two in Polis, but itās actually one lump, which I thought would be a limitation of PCA.
- As a comparative ease of use, I UMAPed the data used in the āPolis-like Visualization of the 2022 House of Councillors Electionā.
- UMAP visibility for the 2022 Upper House election
- Overwhelmingly and clearly divided.
- I did the same thing with 110,000 data and it wasnāt very interesting: 110,000 UMAP.
- This, too, well, I think the structure is such that it cannot be reduced to two dimensions even with UAMP.
- 11/4 2024 Junior Researchersā Report Meeting.
What to do this week
- Iāve done a lot of broad listening from opinion vectors, so I want to do broad listening from text.
- However, now that we have the final data for the poll map, Iād like to analyze that as well.
- Itās Wednesday now, so thereās not much of the week left.
This page is auto-translated from /nishio/é±čØ2024-11-06~2024-11-10 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. Iām very happy to spread my thought to non-Japanese readers.