Skip to content

README.md

Latest commit

 

History

History
239 lines (180 loc) · 7.07 KB

README.md

File metadata and controls

239 lines (180 loc) · 7.07 KB

Taste Engine 📊

Real-time consumer intelligence platform tracking fashion, brands, and culture trends.

What it does

Tracks trending topics across:

  • Social Media: TikTok hashtags, Twitter engagement, Reddit sentiment
  • Commerce: StockX resale prices, volume tracking
  • Institutional: Fashion Week runways, Super Bowl ads, designer moves
  • Fashion aesthetics: Mob wife, quiet luxury, gorpcore, opiumcore
  • Brands: Chrome Hearts, Rick Owens, Balenciaga, Arc'teryx
  • Footwear: Sambas, Salomon, New Balance

Features

✅ Multi-Platform Tracking

  • TikTok: Hashtag velocity, sound-to-fashion correlation, creator influence
  • Twitter: Real-time engagement metrics, viral moment detection
  • Reddit: Sentiment analysis, community validation
  • StockX: Resale price movements, volume tracking
  • Runway: Fashion Week trends, runway-to-street gap analysis
  • Super Bowl: Major ad campaign tracking, brand spend analysis

✅ Smart Analytics

  • Cross-platform correlation detection
  • Trend velocity scoring (0-100)
  • Predictive insights ("This will peak in 7-14 days")
  • Runway vs reality gap analysis
  • Ad spend vs social reality comparison

✅ Automated Workflows

  • GitHub Actions runs every 3 hours
  • Auto-scans all data sources
  • Generates data-driven posts
  • Auto-posts to Twitter
  • Commits updated data to repo

✅ Data-Driven Posts Real insights like:

"BMW spending $10.5M to revive quiet luxury in the Super Bowl. Trend score: 8/100. The money already wasted."

"Pattern detected: Fur coats 78% street adoption before Fashion Week validated. Culture leads, fashion follows."

Tech Stack

  • Python 3
  • bird CLI for Twitter data
  • GitHub Actions for automation
  • Multi-source data aggregation
  • Real-time trend scoring

Quick Start

1. Clone the repo

git clone https://github.com/codebyellalesperance/taste-analytics.git
cd taste-analytics

2. Set up credentials (IMPORTANT 🔒)

Never commit secrets to git!

For local development:

# Copy the template
cp .env.example .env

# Edit .env and add your credentials
# Get Twitter cookies from your browser after logging in

For GitHub Actions:

  • Go to Settings → Secrets and variables → Actions
  • Add TWITTER_AUTH_TOKEN and TWITTER_CT0
  • See AUTOMATION.md for full setup

3. Run scripts

# Install dependencies
pip install requests

# Run individual collectors
python3 scripts/collect_tiktok.py
python3 scripts/collect_twitter.py
python3 scripts/collect_stockx.py

# Run full analysis
python3 scripts/ultimate_dashboard.py

# Generate posts
python3 scripts/generate_posts.py

Automation

The repo includes GitHub Actions workflows that run automatically:

  • Every 3 hours: Full scan + auto-post 3 tweets
  • Manual: Post on-demand from Actions tab

See AUTOMATION.md for complete setup guide.

Security

🔒 Never commit API keys, tokens, or credentials to git.

  • Credentials go in .env (gitignored) or GitHub Secrets
  • See SECURITY.md for best practices
  • .env.example shows the template

Sample Output

Trend Scores

📈 TREND SCORES (0-100):
opiumcore       [████████░░░░░░░░░░░░]  40/100
archivefashion  [████████░░░░░░░░░░░░]  40/100
chrome hearts   [█░░░░░░░░░░░░░░░░░░░]   6/100

Cross-Platform Insights

🧠 CROSS-PLATFORM INSIGHTS:
1. Audio trend alert: 'Femininomenon by Chappell Roan' with 234,000 uses 
   directly driving coquette aesthetic. Music is the new fashion marketing.

2. PREDICTION: #archivefashion will peak in 7-14 days. Currently at 34M views 
   with +567% growth. Early movers should exit soon.

3. DEATH WATCH: #quietluxury down -67% on TikTok. The algorithm has moved on. 
   Brands still pushing this are already late.

Generated Posts

1. Trend scores right now: opiumcore (40/100), archivefashion (40/100), 
   gorpcore (9/100). The algorithm has spoken.

2. Extreme Shoulders on 27 runways. Street adoption: 3%. Balenciaga's $4,000 
   jackets about to hit clearance.

3. Platform breakdown for 'opiumcore': TikTok (explosive), StockX (rising), 
   Reddit (positive). Triple confirmation = real trend.

Architecture

Data Sources                Analysis                Output
━━━━━━━━━━━━━━            ━━━━━━━━━━━━━━         ━━━━━━━━━━━━━━
TikTok                     Trend Scoring           Twitter Posts
Twitter          →         Correlation    →        Data Commits
Reddit                     Predictions             Artifacts
StockX                     Gap Analysis
Runway
Super Bowl

Project Structure

taste-analytics/
├── .github/workflows/     # Automation workflows
├── scripts/               # Data collectors & analyzers
│   ├── collect_tiktok.py
│   ├── collect_twitter.py
│   ├── collect_stockx.py
│   ├── collect_reddit.py
│   ├── collect_runway.py
│   ├── collect_superbowl.py
│   ├── dashboard.py
│   ├── generate_posts.py
│   ├── master_analyzer.py
│   └── ultimate_dashboard.py
├── data/                  # Generated data (gitignored)
├── output/                # Generated posts (gitignored)
├── .env.example           # Credential template
├── AUTOMATION.md          # Automation setup guide
├── SECURITY.md            # Security best practices
└── README.md              # This file

Use Cases

For Content Creators (@tasteengine)

  • Auto-post data-driven trend insights
  • Stay ahead of mainstream coverage
  • Build authority with real metrics

For Brands (B2B SaaS potential)

  • See what Fashion Week got wrong
  • Know which Super Bowl ads will flop
  • Predict trend peaks before they happen
  • Never waste money on dead trends

The pitch:

"We would have saved BMW $10.5M. We would have told Balenciaga to skip extreme shoulders. We saw mob wife before Vogue. Pay us $5K/month to never waste money again."

YC Application Potential

One-liner:
"Bloomberg Terminal for consumer culture. We tell brands what's going to be cool before it's cool."

Why now:

  • AI can finally process culture at scale
  • Brands desperate for TikTok-speed insights
  • Death of cookies = need new intelligence

Business model:

  • Free: @tasteengine Twitter content
  • $5K/month: Real-time dashboard for brands
  • 100 brands = $6M ARR

Contributing

This is currently a private project. If you have access:

  1. Never commit secrets (use .env or GitHub Secrets)
  2. Follow the security guidelines in SECURITY.md
  3. Run scripts locally before pushing
  4. Test automation workflows before enabling

Cost

GitHub Actions: Free (2,000 min/month, we use ~480)
Data sources: Free (public APIs and scraping)
Total: $0/month ✅

License

Private - All rights reserved

Built by @tasteengine | Follow for real-time trend updates
Questions? Open an issue or check the docs: AUTOMATION.md | SECURITY.md