Projects

Personal OS

In Operation

A personal system for decision support and experience reuse, continuously improved through real-life operation

PythonStreamlitデータ設計意思決定支援

This project is intended for personal use, so interactive demos and source code are not publicly available.

Gallery

Overview of main screens and features.

Overview

A personal system for decision support and experience reuse, continuously improved through actual daily use.

Not designed as a simple recording or management tool, but built to accumulate recurring decisions from daily life and work in a processable and reusable format.

Currently operating primarily for household financial management and experience management, with ongoing review of design and decision criteria through actual use.

Responsibilities

Handled all phases personally: requirements gathering, data design, logic implementation, visualization design, and operational rule organization.

Designed for actual operation, emphasizing a configuration that doesn't excessively increase input burden or decision costs, while continuously making improvements.

Design & Operation Highlights

For this project, "not trying to manage everything" was the first design decision.

The more precise input and complete coverage you require, the more likely operations are to break down, so prioritized a configuration that enables decisions with minimal input.

Focused not on recording itself, but on supporting "how to make the next decision," accumulating experiences and failures in a reusable format.

This system has no completed form, positioned as a long-term evolving personal system that updates the decision criteria themselves through actual use.

Tech Stack

フロントエンド

  • Streamlit

バックエンド

  • Python

データ管理

  • JSON (file-based)

データ処理

  • Python scripts

状態管理

  • Snapshot isolation
  • Monthly file management

実行環境

  • Local

Technical Points & Design Decisions

  • Emphasis on robust data structure and consistent decision criteria for long-term operation

  • Data separated by month and item so that modifying current state doesn't affect past records

  • UI display and report generation use the same data model and logic to prevent inconsistent decision results

  • Visualization prioritizes quick understanding of status and numbers over decoration, adopting layouts that reduce cognitive load