Skip to content

AndyZHENG0715/HKOSignal8TransparencyPortal

BranchesTags

Repository files navigation

HKO Signal 8 Transparency Portal

A public-facing bilingual (English & Traditional Chinese) web portal that independently verifies Hong Kong Observatory's Tropical Cyclone Warning Signal No. 8 issuance against actual observed wind data.

Project Goal: Compare official T8 signal issuance with real wind observations to classify each typhoon event into one of three tiers, measuring early-warning effectiveness and public safety impact.

Overview

The Hong Kong Observatory issues Tropical Cyclone Warning Signal No. 8 when sustained gale-force winds (≥63 km/h) are recorded across the reference network. This portal uses a 3-tier classification system to evaluate whether each signal was issued responsibly—providing sufficient early warning for public preparation while avoiding unnecessary disruptions.

Key Questions This Portal Answers

  • Was the T8 signal justified? (Tier 1/2 = observed sustained gales; Tier 3 = unverified by wind data)
  • How much advance notice did people have? (Early warning minutes between official issuance and actual gales)
  • What was the real-world impact? (Casualties, property damage, and correlation with early-warning lead time)

Data & Methodology

Raw Data Source

  • Hong Kong Observatory 10-minute mean wind observations from 8 reference anemometer stations across Hong Kong
  • Timeframe: 2022–2025 (11 typhoon events)
  • Reference Network (HKO official):
    • Cheung Chau (長洲)
    • Chek Lap Kok (赤鱲角)
    • Kai Tak (啟德)
    • Lau Fau Shan (流浮山)
    • Sai Kung (西貢)
    • Sha Tin (沙田)
    • Ta Kwu Ling (打鼓嶺)
    • Tsing Yi (青衣)

3-Tier Classification System

Tier Condition Indicator
Tier 1 ≥4 stations ≥63 km/h for ≥3 consecutive 10-min intervals (~30 min) 🟢 Sustained T8 Wind Speed Verified
Tier 2 Wind-lull-wind pattern (gales → drop → gales re-emerge) ⚠️ Reappear T8 Wind Speed Verified
Tier 3 Neither Tier 1 persistence nor Tier 2 pattern detected ❓ Unverified
No Signal No official T8 issued ➖ Not Applicable

Key Insight: Tier 1 & 2 demonstrate that HKO's "expected to persist" criterion was met by observation data. Tier 3 signals require HKO to provide public explanation.

Project Structure

HKOSignal8TransparencyPortal/
├── README.md                          # This file
├── HKO Signal 8 Transparency Portal PRD.md  # Full specification
├── data/
│   ├── index.json                    # Summary of all 11 events
│   └── events/                       # Per-event detailed payloads
│       ├── chaba.json
│       ├── koinu.json
│       ├── ma-on.json
│       ├── nalgae.json
│       ├── ragasa.json
│       ├── saola.json
│       ├── talim.json
│       ├── tapah.json
│       ├── toraji.json
│       ├── wipha.json
│       └── yagi.json
├── scripts/
│   ├── build_event_data.py           # CSV → JSON pipeline
│   └── README.md                     # Usage instructions
├── prototypes/
│   ├── charts.html                   # Chart visualizations
│   ├── poster.html                   # Project poster (HTML)
│   ├── QR Code to GitHub Repo.png    # QR code for easy repo access
│   └── README.md                     # Chart documentation
├── typhoon_data/
│   ├── time_of_signal_8.md           # Official signal timings (metadata)
│   ├── casualty_and_lost_of_signal_8.md  # Casualty & property data
│   └── <Event>/                      # Raw HKO CSV files per typhoon
│       └── *.csv                     # 10-minute wind observations
├── Group Project Report & Poster/
│   └── Group Project Report & Poster.txt  # Links to report & poster
└── assets/                            # Images, logos, etc.

Getting Started

Prerequisites

  • Python 3.10+ (only standard library modules used)
  • Modern web browser (ES6+ support, no IE11)
  • Static HTTP server (for local preview)

Quick Start

  1. Clone or download this repository

    git clone <repo-url>
cd HKOSignal8TransparencyPortal

  2. Generate JSON data from raw CSVs (if updating events)

    python scripts/build_event_data.py

  3. Preview the portal locally

    # Option 1: Python
python -m http.server 8000

# Option 2: Node
npx serve

  4. Open in browser

    • Charts: http://localhost:8000/prototypes/charts.html

Data Pipeline

Raw CSV Files (typhoon_data/EVENT/*.csv)
         ↓
   build_event_data.py (Python script)
         ↓
Generated JSON (data/index.json + data/events/*.json)
         ↓
   Web Portal & Charts

Adding a New Typhoon Event

  1. Collect raw CSVs from HKO and place them in a new folder under typhoon_data/

    • Folder naming convention: <EventName>_<Date> (e.g., Nova_20251010)
    • CSV format: yyyymmdd-hhmm-latest_10min_wind.csv

  2. Update metadata files:

    • Add official T8/T10 signal timings to typhoon_data/time_of_signal_8.md
    • Add casualty & property loss data (if available) to typhoon_data/casualty_and_lost_of_signal_8.md

  3. Regenerate data

    python scripts/build_event_data.py
    • Script auto-detects new event folder
    • Computes tiers deterministically
    • Outputs updated data/index.json and data/events/<event>.json

  4. Verify results in the portal

Current Events (11 Typhoons)

Event Year Severity Tier Early Warning Casualties Notes
Chaba 2022 T8 Tier 3 — 3 injured Unverified
Ma-on 2022 T8 Tier 3 — 2 injured Unverified
Nalgae 2022 T8 Tier 3 — 3 injured Unverified
Talim 2023 T8 Tier 3 — 9 injured Unverified
Saola 2023 T10 Tier 2 1,060 min 86 injured Reappear pattern verified
Koinu 2023 T8 Tier 3 — 29 injured Unverified
Yagi 2024 T8 Tier 3 — 9 injured Unverified
Toraji 2024 T8 Tier 3 — 1 injured Unverified
Wipha 2025 T10 Tier 3 — TBD Unverified
Tapah 2025 T8 Tier 3 — TBD Unverified
Ragasa 2025 T10 Tier 1 720 min TBD Sustained verified

Portal Features (Planned/In Progress)

Pages

  • Homepage – Key metrics, event cards (tier-based colors), bilingual toggle
  • Event Detail – Per-event timeline, station statistics, FAQ
  • Comparison Tool – Multi-event analysis, side-by-side metrics
  • Methodology – Tier definitions, HKO vs algorithm comparison, educational FAQs

Charts

  1. Tier Compliance Pie – Share of verified (Tier 1+2) vs unverified (Tier 3) signals
  2. Early Warning Lead/Lag Bar – Minutes between official issuance and observed gales
  3. Casualty/Property Scatter – Correlation between early warning and real-world impact
  4. Event Timeline Heatmap – 10-min timeline × 8 stations, color-coded by wind speed
  5. Per-Event Interactive Timeline – Station counts with T8 threshold reference line

Design

  • Bilingual UI (English & Traditional Chinese 繁體中文)
  • Mobile-first responsive (breakpoint ~768px)
  • Accessible color palette – Clear distinction between tier indicators
  • Interactive charts – Chart.js 4.x, responsive, touch-friendly on mobile

Technical Stack

  • Frontend: Vanilla HTML5/CSS3 + JavaScript (ES6+)
  • Charts: Chart.js 4.x (via CDN)
  • Data Format: JSON (pre-generated, no backend API)
  • Data Processing: Python 3.10+ (standard library only)
  • Deployment: Static site (GitHub Pages compatible)
  • Browser Support: Modern browsers (no IE11)

Bilingual Support

Translations use a centralized JSON dictionary (i18n/) with language-specific keys. Components toggle between English and Traditional Chinese via a language selector.

Example:

<h2 data-key="timelineComparison">Timeline Comparison</h2>

Both keys are resolved at render time based on the active language context.

Success Criteria

✅ Display all 11 typhoon events with automated tier classification (no hardcoding) ✅ Calculate early warning time (Official start → Algorithm detection) ✅ Bilingual interface with seamless language toggle ✅ Interactive per-event timeline charts ✅ Multi-event comparison tool ✅ Tier-based visual distinction (colors, icons, borders) ✅ Reference network strictly adheres to 8 HKO stations ✅ Neutral educational framing (no subjective judgments)

Key Design Principles

  1. Public-Facing: No jargon—assumes audience has no meteorology background
  2. Neutral Tone: Use "Tier 1/2/3" classification rather than "appropriate/inappropriate"
  3. Transparent: Explain both benefits (early warning saves lives) and costs (unnecessary signals cause economic loss)
  4. Data-Driven: All claims backed by actual HKO observation records
  5. Maintenance-Free: Once CSV/metadata are updated, one script run regenerates all JSON

Project History

  • Project Goal: Evaluate transparency of HKO Signal 8 issuance
  • Initiative: Group project for Hong Kong Baptist University
  • Data Span: 11 typhoon events (2022–2025)
  • Status: Data collection & analysis complete; portal UI in development

Contributing & Issues

Reporting a Bug

Please check the raw CSV files and metadata markdown files for accuracy. If a calculation or chart rendering is incorrect, please file a detailed issue with:

  • Event name and date
  • Expected vs actual tier/early-warning minutes
  • Screenshot or error log

Adding Events

Follow the "Adding a New Typhoon Event" section above.

Updating Visualizations

Chart prototypes are in prototypes/charts.html. Modify chart types, colors, or labels there and test locally before committing.

License

This project is licensed under the MIT License – see LICENSE file for details.

MIT License is permissive and free for educational, research, and commercial use. You are free to use, modify, and distribute this project with minimal restrictions.

Acknowledgments

  • Data Source: Hong Kong Observatory (HKO) 10-minute mean wind observations
  • Reference Standard: HKO Tropical Cyclone Warning System (effective 2007–present)
  • Project Sponsor: Hong Kong Baptist University

Contact & Questions

For questions about the methodology, data accuracy, or to request new features:

  • See HKO Signal 8 Transparency Portal PRD.md for full specification
  • Review scripts/README.md for data pipeline details
  • Check prototypes/README.md for chart documentation

Last Updated: December 2025

About

No description, website, or topics provided.

Resources

Readme

License

MIT license
Activity

Stars

2 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages