NISHIO Hirokazu[English][日本語]

Diary 2025-06-20

Be aware that words tend to be harsh when you're short on capacity.

nishio oooh, over 1000 people registered! image nishio If we try to distribute what we want more than 1000 people to do, we usually don't have time to adjust for each individual, so we tend to assign a uniform amount. However, with this method, each individual can choose what he or she wants to do from a list of options, so many things that he or she is good at can be done. nishio You can increase the score for important tasks, decrease the score for tasks that have decreased in demand, and add tasks when you have new things you want done. Not all tasks can be done this way, but it's a very good way to do it with 1000+ supporters of a system of tasks that you want done as many as possible. nishio Talking about this

nishio: A social experiment is underway to gamify politics on a scale never seen before. You only have about a month left to participate in this first experiment, so I urge you to experience it for yourself. This development itself is open source, and is being done on a daily basis. Team Mirai Action Board

I feel like I should write a technical article for Qiita.

  • I thought it would be better than writing in the Note

  • Technology Supporting Donation Visualization

    • Visualize data from Stripe API
      • If we give people access to the Stripe admin screen, we can see who has donated, so we have to limit who can access it.
    • The implementation itself was easy, I asked ChatGPT o3 to write it in Python, and he wrote it.
    • I did some trial and error on my end, then put it on GitHub and told Devin, "I can get data with this, so I'll use it as a reference.
    • Skills to clarify unclear requirements
      • What is the data source?
        • I thought it was just taking data from one Stripe account, but I was wrong.
          • To begin with, there's an account for a private political organization created in 2024 and an account for a political party created in 2025.
          • And in some cases, bank transfers are made instead of through Stripe.
      • Output to?
        • I had a requirement to write in Google Spreadsheet and visualize in Looker Studio, and a requirement to create a GET API to display on a website.
        • This is incompatible, and the data you want to display may be different in the first place.
      • Conclusion.
        • Google Spreadsheet(A) to book bank transfers
        • GitHub Actions to read and combine two Stripe accounts and (A)
        • Output time-series data summarized by day to Google Spreadsheet, which is then visualized in Looker Studio
        • For the matter of GET API for websites, we decided to generate JSON and upload it to Gist.
    • Devin is very active
      • As for the website, I just said, "Hit the API and display it," and gave them the API and an image of the screen design, and it usually worked.
      • Finally, a little human conditioning was needed.
      • This fits the utilization pattern of the Devin video, "a human does the beginning and end, and Devin does the middle.

  • Technology Supporting Supporter Geographic Distribution Visualization

    • If you take detailed addresses, you'll have to deal with them as personal information, which is more burdensome and not what you want in the first place.
    • But I'd like some information on "roughly where they are."
    • →Postal Code
    • Initial Configuration
      • Volunteer Registration with Google Form
      • Enter Google Spreadsheet
      • Google App Script to hit the API and turn a zip code into an address.
    • Changing requirements
      • Action Board Release
      • I want to eliminate duplicates and add up people who registered on Google Form and people who registered only on Team Mirai Action Board.
        • Asking people who have entered the action board to fill out the Google Form again is subtle.
        • Different people prefer different entrances and want to maintain both.
      • Action board data is in Supabase.
        • A bit loaded to handle with Google App Script
        • I decided to move my entire data analysis pipeline to GitHub Actions.

  • It was completed and put into Looker Studio a few days before the switchover, but when we switched to production, we noticed the problem

    • Isn't it strange that there is a hole in Osaka City?
    • I shook it to Devin in a messy way and it was destroyed.

  • Visualization of the degree of Devin utilization?

  • Technology to Support Policy PR

    • data collection
    • Devin

who wants to talk to a human

I think it's a no-brainer to save the cost of screening the wrong people to get them in first and then ban them if they're inappropriate.

This page is auto-translated from /nishio/日記2025-06-20 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.

(C)NISHIO Hirokazu / Converted from Markdown (en)
Source: [GitHub] / [Scrapbox]