An interactive geospatial intelligence tool designed to help H-1B applicants navigate the new FY 2027 Weighted Selection Rule. It combines official Department of Labor (DOL) wage data with a massive database of historical company filings.

🔴 Live Demo

Starting March 2026, the US Department of Homeland Security (DHS) is shifting the H-1B lottery from a random selection to a Wage-Level Weighted Selection.

• Level 4 (Senior) roles will have significantly higher odds (+107%).
• Level 1 (Entry) roles will have significantly lower odds (-48%).

Applicants currently have to blindly guess which job titles or locations yield the best odds. This tool solves that opacity.

• Real-time Odds Calculation: Visualizes "Safety Levels" (Green/Safe vs. Red/Risky) across 3,200+ US counties.
• Smart Zoom: Instantly drill down from State to County level.
• Hybrid/Remote Logic: Built-in compliance alerts for "Lowest Wage Rule" scenarios.

• Historical Data: Search 200,000+ past H-1B filings (FY 2022–2025) to see exactly what SOC codes companies like Google, Amazon, or Tesla actually use.
• O*NET Mapping: Automatically links internal job titles (e.g., "BI Analyst") to official government SOC codes (e.g., "Data Scientist").

• Resume Analysis: Uses Google Gemini / Llama 3 to analyze user job descriptions.
• Reverse Mapping: Matches natural language descriptions to the closest official O*NET Occupation Code using Retrieval-Augmented Generation (RAG).

This project uses a high-performance Serverless & Static architecture to handle massive datasets without a backend database.

Instead of querying a slow SQL database, the 200,000+ record LCA dataset is pre-processed into static JSON shards.

• Process: Python ETL pipeline (process_lca.py) splits the data into tiny chunks based on company prefixes (e.g., db/GO.json for Google).
• Result: O(1) lookup time. When a user types "Go", the app fetches only the relevant ~150KB shard, ensuring instant search results on mobile/edge networks.

To render 3,000+ counties instantly:

• Pre-Computation: Wage data is pre-calculated per SOC code during build time.
• Runtime: The app fetches lightweight GeoJSON vector tiles and "paints" the wage data onto the map layer directly in the browser using mapbox-gl, avoiding heavy server-side rendering.

• Frontend: Next.js 14 (App Router), TypeScript, Tailwind CSS
• Maps: Mapbox GL JS (react-map-gl), Turf.js (Geospatial analysis)
• Data Pipeline: Python (Pandas) for ETL, Static JSON Sharding
• AI: Google Generative AI SDK (Gemini)
• Infrastructure: Vercel (Edge Network)

• Node.js 18+
• Mapbox API Token (Free tier)
• Google Gemini API Key

1. Clone the repository

   git clone [https://github.com/idealidler/h1b-wage-map.git](https://github.com/idealidler/h1b-wage-map.git)
   cd h1b-wage-map

2. Install dependencies

   npm install

3. Set up Environment Variables Create a .env.local file:

   NEXT_PUBLIC_MAPBOX_TOKEN=pk.your_mapbox_token_here
   GEMINI_API_KEY=your_google_ai_key_here

4. Run the Data Pipeline (Optional) If you want to re-process the raw DOL data:

   cd data-pipeline
   python3 clean_data.py  # Generates Wage Map JSONs
   python3 process_lca.py # Generates Company DB Shards

5. Run the App

   npm run dev

This tool is for informational purposes only and does not constitute legal advice.

• Wage Data: Official US Department of Labor (DOL) FLC Data Center.
  • Current Dataset: July 2025 – June 2026 (Active Cycle).
• LCA Data: OFLC Historical Disclosure Data (FY 2022 – 2025).
• Data Updates:
  • Wage Levels: Updated Annually (July).
  • Company Filings: Updated Quarterly.

Akshay Jain

• LinkedIn: @akshayjain128
• GitHub: @idealidler

Built with ☕️ and 💻 to support the immigrant community.