Big Tech Takes Dashboard

A prediction market dashboard tracking Polymarket predictions for major technology companies, providing real-time insights into market sentiment about tech company events and leadership.

Features

• Multi-Company Tracking: View prediction markets for Apple, Google, Microsoft, Amazon, Meta, Tesla, and NVIDIA
• Enhanced Stock Metrics: Comprehensive daily data including current price, daily change %, day's high, and day's low
• Market Intelligence: Display active and resolved markets with yes/no probabilities, trading volume, and resolution dates
• Color-Coded Performance: Green for gains, red for losses with automatic calculation
• Hybrid Data Fetching: Intelligent API strategy that balances speed and coverage
• Automatic Daily Refresh: Data updates once per day via Vercel Cron
• Smart Caching: 5-minute page revalidation for fresher data, cache-busting on manual refresh
• Responsive Design: Mobile-first carousel navigation and adaptive grid layouts
• Admin-Only Manual Refresh: Secure manual data refresh with environment variable control and custom dialog

Setup Instructions

1. Environment Variables

Copy .env.example to .env.local and fill in the required values:

cp .env.example .env.local

DATABASE_URL (Required)

• PostgreSQL connection string from Neon
• Format: postgresql://user:password@host/dbname?sslmode=require

FINNHUB_API_KEY (Required for stock prices)

• Get a free API key at: https://finnhub.io/dashboard
• Free tier: 60 requests per minute (more than sufficient for daily refresh)

CRON_SECRET (Required for data refresh)

• Used to protect the refresh endpoint
• Generate a secure random string: openssl rand -hex 32
• Must be set both locally and in Vercel environment variables

MANUAL_DATA_REFRESH (Optional - for admin features)

• Controls visibility of manual data refresh button
• Set to 'true' to enable admin-only manual refresh feature
• When false or unset, refresh button is hidden and API endpoint returns 403
• Provides defense-in-depth security (frontend + backend enforcement)

2. Database Setup

The database schema includes three main tables:

• companies: Tech companies being tracked
• stock_data: Current stock prices with comprehensive daily metrics (high, low, open, previous_close)
• polymarket_predictions: Prediction markets with event slug support

Create the database schema:

psql $DATABASE_URL -f scripts/create_schema.sql

This script creates all tables, indexes, and inserts the initial set of tracked companies.

Optional: Clear predictions (when needed):

psql $DATABASE_URL -f scripts/clear_predictions.sql

Use this to remove all prediction market data and force a fresh fetch. Useful when:

• Testing changes to market filters
• Removing stale or duplicate markets
• Updating search logic

3. Initial Data Load

Once environment variables are configured:

Option A: Use the Dashboard UI (Requires MANUAL_DATA_REFRESH=true)

1. Set environment variable: MANUAL_DATA_REFRESH=true
2. Start your dev server: npm run dev
3. Click the "Refresh Data" button in the dashboard header
4. Enter your refresh key in the dialog
5. Click OK and wait 5-10 seconds for data to load

Option B: Use cURL

curl -X POST http://localhost:3000/api/refresh \
  -H "Authorization: Bearer $CRON_SECRET"

This will fetch:

• Comprehensive stock data from Finnhub API (price, high, low, open, previous close) with 1-hour caching
• Prediction markets from Polymarket Gamma API (using hybrid events + pagination approach)

Note: After manual refresh, the page automatically reloads with cache-busting to show fresh data immediately. Pages revalidate every 5 minutes for automatic freshness.

4. Automatic Updates

The Vercel Cron job runs once daily at midnight UTC to automatically refresh all data.

Cron Schedule: 0 0 * * * (runs at 00:00 UTC) Frequency: Once per day (Vercel free plan limit) Expected Load: ~70 API calls per refresh (well within free tier limits)

API Endpoints

• GET /api/companies - List all active companies with stock data
• GET /api/markets/[ticker] - Get prediction markets for a specific company (limit 25)
• POST /api/refresh - Trigger manual data refresh (requires Bearer token authorization)

Tech Stack

• Frontend: Next.js 16 with App Router, React 19, Server Components
• Database: Neon PostgreSQL with Vercel Postgres SDK
• APIs:
  • Finnhub Stock API (free tier: 60 calls/minute, comprehensive quote data)
  • Polymarket Gamma API (no rate limits)
• Scheduling: Vercel Cron (daily refresh at midnight UTC)
• Styling: Tailwind CSS v4.1.9 with OKLCH color space and Polymarket branding
• UI Components: shadcn/ui (Radix UI primitives) + Embla Carousel
• Deployment: Vercel with ISR (5-minute revalidation)

UI Features

Enhanced Tab Navigation

Company tabs display comprehensive information:

• Full company name (e.g., "Meta (Facebook)")
• Stock ticker (e.g., "META")
• Current stock price (e.g., "$620.80")

Benefits: Users see all stock prices at a glance without switching tabs.

Compact Stock Metrics Strip

Replaced the large stock card with a responsive 4-column grid:

Current Price	Daily Change	Day's High	Day's Low
$438.57	+2.45%	$440.50	$435.20

Features:

• Color-coded daily change (green for positive, red for negative)
• Automatic percentage calculation from previous close
• Responsive: 2 columns on mobile, 4 columns on desktop
• Subtle background with optimal spacing

Architecture Highlights

Hybrid Data Fetching Strategy

The dashboard uses an intelligent hybrid approach to fetch Polymarket predictions:

1. Events-Based Primary (7-10 API calls):

   • Efficient for companies with dedicated event pages (Apple, Google, Tesla, NVIDIA)
   • Searches through events by company name, extracts markets
   • ~5-7x faster than pagination

2. Pagination Fallback (50 API calls):

   • Reliable for companies with scattered markets (Microsoft, Amazon, Meta)
   • Searches through individual markets in batches of 100
   • Guarantees coverage even when events-based fails

3. Max Selection Logic:

   • If events-based finds <5 markets, trigger pagination
   • Return whichever approach found MORE markets
   • Result: 31% fewer API calls with 25-75% better coverage

Smart Filtering

• Alternate Search Terms: Company-specific keywords (e.g., "gemini" for Google, "iphone" for Apple) capture product-specific markets
• Word Boundary Matching: Prevents false positives (e.g., "Amazon" won't match "Amazonia")
• Stale Data Cleanup: Automatically removes markets >7 days old

Event Slug Support

Polymarket groups related markets under shared event pages (e.g., "Next CEO of Apple?" contains multiple candidates). The dashboard:

• Tracks parent event_slug for grouped markets
• Uses event_slug when present, falls back to market_slug for standalone markets
• Ensures all prediction links navigate to the correct Polymarket page

Development

# Install dependencies
npm install

# Run development server
npm run dev

# Build for production
npm run build

# Start production server
npm start

Documentation

• API Implementation: /documentation/apiImplementation.md - Comprehensive guide to data fetching strategies, enhanced stock metrics, caching, and critical technical decisions
• Design System: /documentation/design.md - UI/UX specifications, Polymarket branding, and component guidelines

Performance

• API Efficiency: 31% fewer calls than baseline pagination approach (~34 calls per company avg)
• Coverage: 25-75% more markets found for major companies (Apple, Google)
• Stock Data: Comprehensive 6-field capture (price, high, low, open, previous_close, market_cap)
• Page Performance: 5-minute ISR revalidation balances freshness with speed
• Cache Strategy: 1-hour stock API caching protects against refresh spam
• Cost: $0/month (within free tier limits for all services)
• Freshness: Daily automated updates at midnight UTC + manual refresh with cache-busting

Contributing

When adding new companies or features:

1. Add company to companies table (update scripts/create_schema.sql or add via SQL INSERT)
2. Add alternate search terms to getAlternateSearchTerms() in /app/api/refresh/route.ts
3. Test using manual refresh before enabling automatic updates
4. Update documentation to reflect changes

License

MIT