An enterprise-grade automation engine leveraging n8n and LLMs to bridge the gap between raw web data and actionable market intelligence.
An autonomous research agent capable of replicating complex human analysis. By orchestrating multi-step workflows, the system transforms raw, fragmented data into structured, executive-ready reports—effectively reclaiming hours of manual labor.
n8n | Google Gemini (LLM) | JavaScript | REST APIs
During my internship, I was tasked with modernizing how we interact with market data. My work resulted in a multi-layered system:
- User Interface: A custom web-interface chatbot allowing non-technical users to query live market data.
- The Logic Engine: A series of data market research workflows to scrape, clean, and analyze competitor trends.
Important
Note on this Repository: Due to proprietary data protections accompanying my contribution to Wayfair, the code provided here is a Fundamental Design Architecture. It showcases the core logic, node structures, and LLM prompting strategies used in the production model, demonstrating how a simple market research workflow can be scaled to enterprise needs.
- Input: Raw URL/Search Query via HTTP Request.
- Orchestration: n8n workflow logic (filtering, JSON parsing).
- Intelligence: Google Gemini (LLM) for sentiment analysis and competitive synthesis.
- Output: Structured PDF/Email report for stakeholders.
- Linkedin: www.linkedin.com/in/kardal-haist-a45a50387
- Custom JavaScript Logic
llm-json-sanitizer.js: Cleans AI formatting.amazon-link-sorter.js: logic to categorize URLs.product-scraper-cheerio.js: DOM parsing for product details.data-aggregator.js: Final ETL merge logic.html-minifier.js: Generates the downloadable report.