Transform opaque government procurement data into searchable, filterable business intelligence -- powered by g0v community APIs and automated analysis pipelines.
Gov Procurement Analytics 將台灣政府採購標案資料轉化為可搜尋、可篩選的情報系統，降低資訊不對稱。適合供應商、業務開發與研究者用於追蹤標案趨勢、競品動態與機會盤點。
Gov Procurement Analytics converts Taiwan government tender data into searchable and filterable business intelligence. It supports opportunity discovery, competitor tracking, and procurement trend analysis.
🏛️ Gov Procurement Analytics 是一套台灣政府採購標案智慧分析系統，將不透明的政府採購資料轉化為可搜尋、可篩選的商業情報。📊 串接 g0v 開放社群 API（pcc-api.openfun.app）進行批次資料下載，支援斷點續傳、速率限制與進度追蹤。🔍 核心篩選引擎運用 80+ 關鍵字橫跨五大專業服務類別——廣告行銷、軟體開發、網頁設計、AI 部署、視覺設計——並自動排除營建、醫療、清潔等不相關領域。🌐 自動生成 31KB 的互動式網頁介面，提供分類篩選、關鍵字搜尋與書籤功能。📌 搭配 Go 語言 + SQLite 建構的書籤伺服器，支援標案收藏、管理與註記。🐍 Python 資料管線提供乾淨的程式化 API，可自訂日期範圍下載、關鍵字搜尋與 CSV 匯出。💡 將原本需要數小時人工搜尋的標案探索流程，縮短至數分鐘自動完成。適合需要追蹤政府標案商機的企業與業務開發團隊。
Taiwan's government procurement market represents billions in annual spending, yet the official portal (web.pcc.gov.tw) makes it nearly impossible to efficiently discover relevant opportunities. Tenders are buried in paginated lists with no meaningful filtering, no category-based search, and no way to match opportunities to a company's actual capabilities.
This project closes that gap. It batch-downloads procurement data via the g0v community API, applies intelligent keyword filtering across service categories (advertising, software development, web design, AI deployment, visual design), excludes irrelevant sectors (construction, medical, cleaning), and generates a browsable web interface with bookmark functionality. What used to take hours of manual searching now runs in minutes.
g0v Community API (pcc-api.openfun.app)
|
Batch Download Engine
(rate-limited, resumable, with progress tracking)
|
+----------+----------+
| |
Raw JSONL Storage Category Splitter
(by date/year) (engineering/goods/services/other)
|
Keyword Filter Engine
(5 service categories, exclusion rules, bid-type filtering)
|
+----------+----------+----------+
| | | |
Category Summary Interactive Bookmark
JSONL Report Web UI Server
Files (Markdown) (HTML/CSS/JS) (Go + SQLite)
| Script | Purpose |
|---|---|
download_pcc_data.py |
Full-featured batch downloader with retry logic, rate limiting, date-range support, keyword search, and CSV export. The PCCDownloader class provides a clean API for custom data retrieval. |
download_2026.py |
Focused downloader for current-year tenders with resumable progress tracking and automatic category splitting. |
filter_tenders.py |
Intelligent tender matching engine. Classifies tenders across 5 service categories using 80+ keywords, applies exclusion rules for irrelevant sectors, and generates categorized output with summary reports. |
split_tenders.py |
Large-dataset organizer. Splits bulk JSONL files by procurement category (engineering, goods, services) and year, with automatic index generation. |
generate_web.py |
Generates a full interactive web application (31KB) with category filters, search, bookmarking, and a responsive gradient UI for browsing filtered tenders. |
bookmark-server/ |
Go-based REST API server with SQLite persistence for saving, managing, and annotating bookmarked tenders. |
| Layer | Technology |
|---|---|
| Data Ingestion | Python, requests (with retry/rate-limit) |
| Data Source | g0v Community API (pcc-api.openfun.app) |
| Data Format | JSONL (line-delimited JSON) |
| Filtering | Python (80+ keywords, 5 categories, exclusion rules) |
| Web Interface | Generated HTML/CSS/JS (responsive, interactive) |
| Bookmark Server | Go, SQLite, net/http |
| UI Testing | Dedicated uitest/ directory |
# Install dependencies
pip install requests
# Download current-year tender data
python download_2026.py
# Filter tenders matching your company's capabilities
python filter_tenders.py
# Generate interactive web browser
python generate_web.py
# Start bookmark server (optional)
cd bookmark-server && go run main.gofrom download_pcc_data import PCCDownloader
downloader = PCCDownloader(output_dir="my_data")
# Download by date range
downloader.download_date_range("20260101", "20260131")
# Search by keyword
downloader.download_all_tenders_by_keyword("software development")
# Export to CSV
data = downloader.search_by_title("cloud services")
downloader.export_to_csv(data, "cloud_tenders.csv")The filter engine matches tenders against these professional service categories:
| Category | Sample Keywords |
|---|---|
| Advertising & Marketing | branding, PR, media, social, SEO, events, campaigns |
| Software Development | software, APP, API, system development, custom modules |
| Web Design | website, RWD, responsive, e-commerce, portal |
| AI Deployment | AI, machine learning, LLM, GPT, automation, algorithms |
| Visual Design | UI/UX, graphic design, multimedia, video, animation |
Automatically excludes: construction, medical, cleaning, security, catering, printing, and other irrelevant sectors.
- Official Source: Government E-Procurement System
- API Provider: g0v Community (open source civic tech)
- Data License: Government open data -- free for personal, educational, and reporting use. Commercial use should reference official open data sets.
Huang Akai (Kai) -- Founder @ Universal FAW Labs | Creative Technologist | Ex-Ogilvy | 15+ years experience
MIT