Research Radar 📡

AI-Powered Market Intelligence Dashboard for Startups & Developers.

Research Radar automatically monitors global research reports (e.g., PwC), uses LLMs to filter out "corporate fluff," and extracts concrete business opportunities and technology gaps relevant to startups.

---

this project is now being refactorized from streamlit website to Vanila HTML/CSS/JS website in order to divide backend and frontend

---

✨ Key Features

• 🎯 Startup-Focused AI Filter:
  • Automatically discards generic macro-economic statements (e.g., "GDP is growing").
  • Keeps & Scores specific pain points, unmet needs, and operational inefficiencies (e.g., "Banks struggle with unstructured data").
• 🇰🇷 Full Korean Language Support:
  • Dual-Language UI: One-click toggle between English and Korean.
  • Content Translation: AI automatically translates extracting signals (Summary, Evidence, Expected Value) and Report Metadata into Korean.
• 📊 Interactive Dashboard:
  • Smart Filters: Filter opportunities by Industry, Technology tag, and Importance Score.
  • Detail View: Click any card to see full evidence, confidence score, and source links in a modal popup.
  • Admin View: Inspect "Discarded Signals" to verify what the AI is filtering out.
• ⚡ Real-time Ingestion:
  • Fetch RSS feeds, parse HTML/PDF, and generate insights on-demand.

🛠️ Tech Stack

• Frontend: Streamlit (Python)
• AI Core: OpenAI GPT-3.5 Turbo (JSON Mode)
• Data Engineering: Feedparser, BeautifulSoup4, PyPDF
• Data Validation: Pydantic
• Storage: JSON (Lightweight MVP)

🚀 Getting Started

1. Installation

# Clone the repository
git clone https://github.com/yourusername/radar.git
cd radar

# Install dependencies
pip install -r requirements.txt

2. Configuration

Create a .env file in the root directory and add your OpenAI Key:

OPENAI_API_KEY=sk-your-api-key-here

3. Running the Dashboard

python -m streamlit run app.py

Visit http://localhost:8501 in your browser.

---

🔄 Managing Data

Reprocess & Translate All Data

If you want to re-run the AI extraction on all existing reports (e.g., to apply new filters or generate Korean translations for old data):

python reprocess_data.py

Reset Data

To completely wipe all ingested data and start fresh:

python reset_data.py

📂 Project Structure

• src/: Core logic (Ingestion, Parsing, LLM, Models).
• data/: JSON storage for Reports and Opportunity Cards.
• app.py: Streamlit Dashboard entry point.
• reprocess_data.py: Script to batch-update existing data.

---

Built for the "Agentic Coding" Project.