Real-time predictive analytics platform that correlates data across four public API domains — weather, economic indicators, news sentiment, and transportation — to surface cross-domain insights via an interactive Flask dashboard with WebSocket updates.
| Domain | Source | Cache TTL | What it provides |
|---|---|---|---|
| Weather | OpenWeatherMap API | 30 min | Current conditions, 5-day forecasts, temperature trends |
| Economic | Alpha Vantage API | 60 min | Market indices, exchange rates, sector performance |
| News/Social | News API | 15 min | Trending topics, sentiment scores, keyword frequency |
| Transportation | TomTom Traffic API | 10 min | Traffic density, congestion indices, transit metrics |
| Capability | Implementation | Detail |
|---|---|---|
| Cross-domain correlation | Pearson + rolling window | Identifies non-obvious relationships between weather, markets, traffic, and news |
| Natural language queries | Intent classification + entity extraction | Ask "How does temperature affect traffic congestion?" and get visual answers |
| Real-time updates | Flask-SocketIO with WebSocket fallback | Dashboard refreshes without page reload when new data arrives |
| Predictive analytics | scikit-learn time-series models | Forecasts future trends using cross-domain feature combinations |
| Graceful degradation | 3-tier fallback: API → cache → demo data | System stays operational even when all external APIs are down |
| Use case templates | 4 pre-built analytical environments | Supply chain, public health, urban infrastructure, financial strategy |
git clone https://github.com/damsolanke/Cross_Domain_Predictive_Analytics_Dashboard.git
cd Cross_Domain_Predictive_Analytics_Dashboard
pip install -r requirements.txt# Configure API keys (optional — runs with demo data without them)
cp .env.example .env
# Edit .env with your API keys
# Start the dashboard
python run.py
# → http://localhost:5000The system works without API keys — it falls back to generated demo data with realistic domain patterns.
| Decision | Why | Tradeoff |
|---|---|---|
| Flask over FastAPI | Jinja templates for server-rendered dashboards, mature SocketIO integration | No async by default; mitigated by eventlet/gevent |
| Socket.IO for real-time | Bidirectional communication, automatic reconnection, room-based broadcasts | Additional server process; simpler than polling |
| Demo data fallback | Portfolio reviewers can evaluate the full system without obtaining 4 API keys | Demo patterns are synthetic; real API keys unlock live data |
| Domain-specific cache TTLs | Weather changes faster than economic indicators — cache durations reflect real-world update frequencies | More complex invalidation logic |
| In-memory storage | No database setup required for reviewers; caching layer handles persistence | Data lost on restart; acceptable for analytics dashboard |
| Multi-domain correlation engine | Cross-domain insights are the differentiator — weather×traffic, sentiment×markets | Correlation ≠ causation; confidence scoring helps |
├── app/ # Main application package
│ ├── api/ # API routes + domain connectors
│ │ └── connectors/ # weather, economic, social, transport
│ ├── data_processing/ # cleaner, correlator, transformer, validator
│ ├── models/ # ML prediction models (weather, economic, transport)
│ ├── nlq/ # Natural language query processor + routes
│ ├── system_integration/ # Pipeline orchestration, socket events, correlation engine
│ ├── visualizations/ # Chart formatters (time-series, correlation, dashboard)
│ ├── static/ # CSS, JavaScript, images
│ └── templates/ # Jinja2 HTML templates
├── tests/ # Integration tests
├── app/tests/ # Unit tests (NLP, API, correlation)
├── docs/ # Architecture + API configuration guides
├── scripts/ # Diagram generation
├── .github/workflows/ci.yml # CI pipeline
├── run.py # Entry point (calls create_app factory)
├── requirements.txt # Python dependencies
├── .env.example # Required environment variables
└── LICENSE # MIT
# Run all tests
python -m pytest tests/ app/tests/ -v
# Unit tests only (NLP processor, correlation engine, NLQ API)
python -m pytest app/tests/ -v --ignore=app/tests/test_browser.py
# Integration tests (system integration, pipeline, alerts)
python -m pytest tests/ -v| Use Case | Challenge | Cross-Domain Approach |
|---|---|---|
| Supply Chain | Anticipate disruptions | Weather events × transportation delays × economic indicators |
| Public Health | Optimize resource allocation | Weather patterns × mobility data × social media health sentiment |
| Urban Infrastructure | Plan maintenance proactively | Traffic density × weather conditions × infrastructure usage |
| Financial Strategy | Identify emerging trends | Economic indicators × news sentiment × social media momentum |
MIT — see LICENSE.