Skip to content

Checkpoint 1

Phachara6609650509 edited this page Sep 28, 2025 · 16 revisions

Checkpoint #1: Foundation & AI Prototype

ช่วงเวลาที่ดำเนินการ

วันที่เริ่มต้น: 12 september 2025 วันที่รายงาน: วันที่ 3 october 2025
ระยะเวลา: 3 สัปดาห์


Updated Project Information

Problem Statement

ปัจจุบันรายการ สถานีประชาชน ของ ThaiPBS ได้รับเรื่องร้องเรียนเกี่ยวกับกลโกงจากประชาชนผ่านช่องทาง Line Official Account (Line OA) เป็นจำนวนมาก ซึ่งข้อความที่ส่งเข้ามามีทั้งความยาว เนื้อหา และรายละเอียดที่หลากหลาย ทำให้แอดมินผู้ดูแลต้องรับภาระในการคัดกรองข้อความจำนวนมาก นำมารวมเรียบเรียงเป็นพารากราฟสำหรับจัดเก็บลงฐานข้อมูลและไฟล์ Excel รวมถึงต้องจัดลำดับความสำคัญของเรื่องราวเพื่อส่งต่อให้นักข่าวนำไปสืบค้นและผลิตรายการ

Stakeholders

  • ผู้ร้องเรียน: ประชาชนที่ต้องการแจ้งปัญหากลโกง
  • แอดมิน Line OA: ผู้รับผิดชอบคัดกรองและประมวลผลข้อความ
  • นักข่าว/ทีมรายการ: ผู้ใช้ข้อมูลที่ได้รับการคัดกรองแล้วเพื่อทำข่าว
  • องค์กร ThaiPBS: เจ้าของระบบและผู้กำหนดนโยบาย

Objectives

  1. ลดเวลาการคัดกรองข้อความร้องเรียนจากแอดมิน
  2. เพิ่มความถูกต้องในการจำแนกประเภทเรื่องร้องเรียน
  3. สร้างระบบสรุปอัตโนมัติสำหรับนักข่าว
  4. รองรับการรับข้อความจำนวนมากได้อย่างมีประสิทธิภาพ

Architecture Diagram

image
ประชาชน → Line OA → API Gateway → Lambda Functions → AI Processing → Database
                                       ↓
Admin Dashboard ← WebSocket ← Queue Processing ← AI Results

Updated Assumptions & Constraints

Assumptions:

  • ผู้ใช้มีการเข้าถึง Line และอินเทอร์เน็ต
  • มีงบประมาณ AWS services เพียงพอ
  • ข้อมูลสามารถใช้เพื่อการนำเสนอข่าวได้ตาม PDPA
  • AI สามารถจำแนกข้อความภาษาไทยได้อย่างมีประสิทธิภาพ

Constraints:

  • งบประมาณและทรัพยากรมีจำกัด
  • ระยะเวลาพัฒนา 1 เทอม
  • รองรับภาษาไทยเป็นหลัก
  • ต้องปฏิบัติตาม PDPA

ภาพแสดงความคืบหน้าจาก GitHub Project

สรุปสถานะ Tasks

  • Total Tasks: 25 tasks
  • Done: 8 tasks (32%)
  • In Progress (Planning): 3 tasks (12%)
  • Planned (Checkpoint #2 / Final): 14 tasks (56%)

การแบ่งงานตาม Role

  • Project Manager: 3 tasks (Project Proposal, Planning, Meeting)
  • Backend Developer: 9 tasks (Database, API, RBAC, AI integration)
  • Frontend Developer: 4 tasks (Dashboard, Login, Main Page, History UI)
  • DevOps Engineer: 5 tasks (CI/CD, SQS, RDS, Infrastructure)
  • Security Lead: Security Lead: 1 task (Cognito RBAC Integration)

Tasks ที่เสร็จแล้ว (Done)

  1. [Task] วิเคราะห์ปัญหาของ Thai PBS - Priority: High
  2. [Task] สร้าง Task - Priority: Medium
  3. [Task] Project Proposal - Priority: Medium
  4. [Task] วางแผนเพื่อหาข้อสรุป Feature - Priority: High
  5. [Task] แบ่งหน้าที่ของสมาชิกในทีม - Priority: Medium
  6. [Task] Meeting - Priority: Medium
  7. [Task] Requirement Gathering & Analysis - Priority: High
  8. [Task] ทำ Architecture Diagram - Priority: High

Tasks กำลังดำเนินการ (In Progress – 3 tasks)

  • [Task] Mockup Data ประวัติการแชท
  • [Task] ออกแบบหมวดหมู่ใน Figma สำหรับ Filter
  • [Task] ออกแบบหน้าดูประวัติการแชท

Tasks ที่อยู่ในแผน (Planned – 14 tasks)

  • Frontend Development:
  • ออกแบบหน้า Login
  • ออกแบบหน้า Dashboard
  • หน้า Main แสดงสถิติ
  • สามารถดูประวัติย้อนหลังได้
  • Backend Development & AI:
  • ทำ API สำหรับดึงข้อมูลหมวดหมู่
  • เก็บข้อมูลจาก Summarize ลง RDS
  • เก็บข้อมูลสถิติใน RDS
  • AI Summarize ได้แม่นยำ
  • Security & Authentication
  • เชื่อมต่อ Cognito
  • ทำ Cognito กับ RBAC
  • ทดสอบการ Login แบบ RBAC
  • DevOps / Cloud
  • ทำ CI/CD กับ GitHub Action> * ทำ Cognito กับ RBAC
  • SQS for LINE Webhook
  • เก็บข้อมูล Summarize ลง RDS

Progress Report with STAR Method

Situation (สถานการณ์)

จากการศึกษาปัญหาของรายการสถานีประชาชน พบว่าแอดมินต้องคัดกรองข้อความร้องเรียนจำนวนมากด้วยตนเอง ขาดระบบที่ช่วยจำแนกประเภทและสรุปข้อมูลอัตโนมัติ ทำให้การประมวลผลล่าช้าและมีความเสี่ยงในการตกหล่น

Task (เป้าหมาย)

จัดทำ project foundation ที่ครบถ้วน รวมถึงการวิเคราะห์ requirements, การออกแบบ architecture, และการตั้งค่า development workflow เพื่อเตรียมความพร้อมสำหรับการพัฒนาระบบ AI-powered complaint management

Action (การลงมือทำ)

1. Project Planning & Analysis

  • Requirement Gathering & Analysis (#30) - วิเคราะห์ความต้องการของ stakeholders
  • Meeting (#29) - ประชุมทีมเพื่อวางแผนการทำงานและกำหนด roles ชัดเจน
  • วิเคราะห์ปัญหาของ Thai PBS (#46) - ศึกษา current state และ pain points

2. System Design & Architecture

  • Architecture Diagram (#35) - ออกแบบ high-level architecture ด้วย AWS serverless
  • ภาพแสดงเนื้อหาร้องเรียน Feature (#33) - กำหนด UI/UX requirements
  • Mockup Data preparation (#31) - เตรียมข้อมูลตัวอย่างสำหรับการทดสอบ

3. Technical Foundation Setup

  • Database Design (#51, #44) - วางแผนการจัดเก็บข้อมูลใน RDS และ summarization
  • API Planning (#50) - ออกแบบ RESTful APIs สำหรับการควบคุมข้อมูล
  • Task Management (#32) - ตั้งค่า GitHub Project board และ issue tracking

4. Team Coordination

  • แบ่งหน้าที่ตาม expertise: Backend (1 คน), Frontend (1 คน), DevOps (2 คน), Security (1 คน), PM (1 คน)
  • กำหนด workflow การทำงานแบบ collaborative
  • ตั้งค่า development environment และ standards

Result (ผลลัพธ์)

Project Structure & Planning

  • Complete Project Board: 25 tasks ถูกจัดหมวดหมู่เป็น 4 milestones ชัดเจน
  • Clear Roadmap: มีแผนการพัอนาที่สอดคล้องกับ Well-Architected Framework
  • Team Alignment: ทุกคนเข้าใจบทบาทและความรับผิดชอบของตนเอง

Requirements & Design Clarity

  • Stakeholder Analysis: วิเคราะห์ความต้องการของผู้ใช้ 4 กลุ่มอย่างครบถ้วน
  • Technical Specification: มี architecture diagram และ technology stack ที่ชัดเจน
  • Success Criteria: กำหนดเกณฑ์วัดผลทั้งด้านธุรกิจและเทคนิค

Development Readiness

  • Foundation Complete: พร้อมเริ่ม implementation phase
  • Risk Mitigation: ระบุและวางแผนจัดการความเสี่ยงที่สำคัญ
  • Resource Planning: จัดสรรทรัพยากรและเวลาอย่างเหมาะสม

Well-Architected Pillars

Operational Excellence

  • Project Management: ใช้ GitHub Projects สำหรับ task tracking และ team collaboration
  • Development Workflow: กำหนด branching strategy และ code review process
  • Documentation: มี comprehensive project documentation ใน Wiki

Security

  • Authentication Planning: วางแผนใช้ Amazon Cognito สำหรับ RBAC
  • Data Protection: ออกแบบการเข้ารหัสข้อมูลตาม PDPA requirements
  • Access Control: กำหนดสิทธิ์การเข้าถึงแยกตาม user roles

Reliability

  • High Availability Design: ใช้ multi-AZ deployment สำหรับ critical components
  • Error Handling: วางแผน graceful degradation และ retry mechanisms
  • Monitoring Strategy: กำหนดแผน CloudWatch monitoring และ alerting

Performance Efficiency

  • Serverless Architecture: เลือกใช้ Lambda และ API Gateway สำหรับ auto-scaling
  • Database Optimization: วางแผน indexing strategy สำหรับ fast query performance
  • Caching Strategy: ออกแบบ caching layers ด้วย DynamoDB และ ElastiCache

Cost Optimization

  • Pay-per-use Model: เลือก serverless services เพื่อลดต้นทุนเมื่อไม่มีการใช้งาน
  • Resource Right-sizing: วางแผนการใช้ resources ให้เหมาะสมกับ workload
  • Cost Monitoring: ตั้งค่า billing alerts และ cost tracking

สิ่งที่ทำเสร็จแล้ว

  • Project Proposal และ requirements analysis
  • Team formation และ role assignment
  • High-level architecture design
  • GitHub Project setup และ task planning
  • Technology stack selection
  • Development workflow establishment

กำลังทำอะไร

  • Database schema design และ implementation
  • AI model research และ prototyping
  • API endpoint development
  • Frontend mockup และ component planning
  • AWS infrastructure setup
  • Authentication system design

ปัญหาที่เจอ

  1. AI Model Selection: ต้องประเมินระหว่าง Amazon Comprehend vs Custom Model
  2. Data Privacy: ต้องศึกษา PDPA compliance อย่างละเอียด
  3. Performance Requirements: ต้องทดสอบ throughput ที่แท้จริงของระบบ

แผนสำหรับ Checkpoint #2

  1. Complete AI prototype และ testing
  2. Implement authentication system
  3. Create functional dashboard
  4. Setup CI/CD pipeline
  5. Performance optimization และ monitoring setup

ความคืบหน้าโดยรวม

Progress: 35% ของ total project scope
การดำเนินงานใน Checkpoint #1 เน้นการวางรากฐานที่แข็งแกร่งสำหรับการพัฒนาระบบ ทีมมีความเข้าใจชัดเจนในเป้าหมาย และพร้อมเริ่ม implementation phase ใน Checkpoint #2

ScamReport

Online Fraud Protection System


Navigation

Project Documentation


Milestones

Phase Status Date
Proposal Complete Aug 24
Checkpoint #1 In Progress Sep 21
Checkpoint #2 Planned Oct 19
Final Planned Nov 16

Clone this wiki locally