Policy Pullic Utilization Project
2025-08-19
Team Mirai Cluster Density Sort
2025-08-06
Team Mirai Hierarchical Clustering
2025-08-04
I tried to put 20K data into Polis and it was too big and gave me an error.
Sampled and put in 1/10th.
PCA hasn't finished in an hour.
Maybe because it's implemented iteratively.
2025-07-31
Discussion with [Colin Megill
nishio For Japanese: I'll post this in Japanese as well for Japanese people.
This is not just using the title of the issues, but the entire text, diffs, etc. LLM extracts the contributor's perceived awareness of 0-N issues from a single overall data set and embeds them in a higher dimensional space.
It is my opinion that AI should have really dug deeper into the data at the stage of interviewing users to find out what issues they are feeling. In this case, we didn't do that at the time of the interview, so we are guessing from the text of the issues obtained after the fact. I think you could also guess from the chat logs. My concern is that the cost would be several times higher and that the users have not been given permission for such usage. By the way, this experiment cost 30USD.
Colin told me that it would be interesting to put it in Polis 2.0, and it certainly sounds interesting, so I'm going to do it.
2025-07-30
Team Mirai Problem Awareness Spreading AI
2025-07-27
nishio Team Mirai's Policy Proposal Repository became a public archive with the end of the Upper House election, so I put the last data collection on it. You should be able to find the data of all PRs here in /prs/. If anyone has any analysis to do, please let me know!
https://github.com/team-mirai-volunteer/pr-data/tree/main
Various analysis tools are included here.
https://github.com/team-mirai-volunteer/policy-pr-hub
The main repository is here, so if you have any missing attribute data, etc., please try to get it from here, and the policy-pr-hub code may be helpful.
https://github.com/team-mirai/policy
2025-07-26 policy repo is now a public archive
= No more updates in the future, so it is easy to handle as a data analysis target.
2025-05-30
https://github.com/team-mirai-volunteer/policy-pr-hub/blob/main/README.md
status quo
The "Talking Manifesto" (initiative taken by the government in the formulation of a policy or bill (rather than by bureaucrats) to promote good governance) worked well and attracted more than 1,700 suggestions for revision.
We want to use this data to create better policies.
Initially tried and tested at https://github.com/team-mirai/random, still working on Github Actions here.
I'd like to organize this so others can tinker with it, but it's currently not done.
I want to keep things organized here.
https://github.com/team-mirai-volunteer/policy-pr-hub
---
This page is auto-translated from /nishio/政策プルリク活用プロジェクト 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.