End-to-end data engineering pipeline that ingests, transforms, and warehouses real-time stock market data — built to demonstrate production-grade streaming, orchestration, and dimensional modeling.
View Live Site · Interactive Dashboard
Tracks the full price history of 10 major US equities (AAPL, AMZN, GOOG, META, MSFT, NFLX, NVDA, TSLA, JPM, DIS) through a fully containerized pipeline:
- Streams live market data via Kafka (producer → broker → consumer)
- Orchestrates 23 Airflow DAGs for ETL, fundamentals, earnings, SEC filings, and macro data
- Warehouses everything in a TimescaleDB star schema (3 dimension + 7 fact tables)
- Visualizes results on a Bloomberg Terminal-style dashboard with Chart.js
| Source | Data | Method |
|---|---|---|
| yfinance | Prices, company info, earnings | Python API |
| SEC EDGAR | 10-K/10-Q financial statements | EdgarTools |
| FRED | Fed funds rate, CPI, GDP, unemployment | fredapi |
| Alpaca | Hourly price bars back to 2016 (free paper-trading account) | REST API |
Explore the platform's output — no setup required:
The landing page features a Bloomberg-style ticker tape and architecture overview. The dashboard displays interactive charts for price trends, fundamentals, earnings surprises, and macro indicators.
7 Docker containers work together:
| Service | Image | Port | Role |
|---|---|---|---|
timescaledb |
timescale/timescaledb:latest-pg14 | 5434 | Star-schema warehouse |
airflow-webserver |
Custom (Dockerfile.airflow) | 8081 | DAG monitoring UI |
airflow-scheduler |
Custom (Dockerfile.airflow) | — | DAG execution engine |
zookeeper |
confluentinc/cp-zookeeper | 2181 | Kafka coordination |
kafka |
confluentinc/cp-kafka | 9092 | Message broker |
kafka-producer |
Custom (Dockerfile.kafka) | — | Market data ingestion |
kafka-consumer |
Custom (Dockerfile.kafka) | — | Kafka → TimescaleDB sink |
| Table | Type | Description |
|---|---|---|
dim_company |
Dimension | Ticker, company name, sector, industry, exchange (SCD Type 2) |
dim_date |
Dimension | Year, quarter, month, day, weekend flag |
dim_macro_indicator |
Dimension | FRED macro series metadata |
fact_stock_price_daily |
Fact | OHLCV data per ticker per day, back to each ticker's first traded day |
fact_stock_price_intraday |
Fact | Hourly regular-session OHLCV bars back to 2016, labelled by bar_interval and source |
fact_stock_price_monthly |
Fact | Aggregated monthly averages and total volume |
fact_company_fundamentals |
Fact | Market cap, PE ratios, dividends, beta |
fact_earnings |
Fact | Quarterly EPS: estimate vs actual, surprise % |
fact_sec_financials |
Fact | SEC filing line items, keyed on the full reporting period so quarterly and year-to-date figures stay distinct |
fact_macro_data |
Fact | Fed funds rate, CPI, unemployment, GDP time series |
Built on TimescaleDB for time-series optimized queries on PostgreSQL 14.
| DAG | Schedule | Purpose |
|---|---|---|
etl_stock_data_<ticker> |
Daily | Per-ticker ETL (one DAG per ticker); first run loads the ticker's full history |
populate_dim_company |
On-demand | Load company dimension table |
populate_dim_date |
On-demand | Generate date dimension (1990-2035) |
fundamentals_daily |
Daily | Fetch company fundamentals |
earnings_weekly |
Weekly | Fetch earnings dates and EPS surprises |
sec_financials_quarterly |
Quarterly | Fetch SEC 10-K/10-Q financial statements |
macro_daily |
Daily | Fetch FRED macro indicators |
intraday_prices_daily |
Daily | Reload hourly bars from Alpaca (skipped if no keys are set) |
monthly_aggregate_dag |
Daily | Recompute monthly price aggregations (daily, so the current month stays current) |
csv_export_dag |
Triggered | Export last 30 days to CSV per ticker |
populate_fundamentals |
On-demand | One-off fundamentals backfill |
populate_earnings |
On-demand | One-off earnings backfill |
populate_sec_financials |
On-demand | One-off SEC filings backfill |
populate_macro_data |
On-demand | One-off FRED macro backfill |
That is 23 DAGs in total: 10 per-ticker ETL DAGs plus the 13 listed here.
- Docker and Docker Compose
- A free FRED API key
# Clone and configure
git clone https://github.com/jayhemnani9910/stock-data-platform.git
cd stock-data-platform
cp .env.example .env # Fill in FRED_API_KEY, EDGAR_IDENTITY, and optionally ALPACA_API_KEY/SECRET
# Start all 7 services
docker compose up -d
# Access Airflow UI
open http://localhost:8081 # credentials from .env
# Initialize dimensions (run these DAGs first)
# 1. populate_dim_company
# 2. populate_dim_date
# 3. etl_stock_data_aapl (or any ticker DAG)
# Query the warehouse
docker exec -it timescaledb psql -U data226 -d stockdw \
-c "SELECT * FROM fact_stock_price_daily ORDER BY date DESC LIMIT 10;"├── Dags/ # Airflow DAG definitions
│ ├── dag_config.py # Shared DAG defaults and ticker loading
│ ├── etl_stock_data_dag.py # Per-ticker extract → transform → load
│ ├── populate_dags.py # Dimension and fact table population
│ ├── fundamentals_dag.py # Daily company fundamentals
│ ├── earnings_dag.py # Weekly earnings data
│ ├── sec_financials_dag.py # Quarterly SEC filings
│ ├── macro_dag.py # Daily macro indicators
│ ├── intraday_dag.py # Daily hourly-bar refresh
│ └── monthly_aggregate_dag.py # Monthly rollups
├── scripts/ # Data population and utilities
│ ├── db_utils.py # Shared DB connection, batch insert, upsert SQL
│ ├── populate_dim_company.py
│ ├── populate_dim_date.py
│ ├── populate_company_fundamentals.py
│ ├── populate_earnings.py
│ ├── populate_stock_price_intraday.py
│ ├── populate_sec_financials.py
│ └── populate_macro_data.py
├── SQL/ # Schema and queries
│ ├── schema.sql # Star schema DDL (TimescaleDB)
│ ├── migrations/ # Re-runnable changes for a live volume
│ └── aggregate_monthly.sql # Monthly rollup query
├── tests/ # Unit tests (pytest)
│ └── unit/ # 202 tests across 10 modules
├── site/ # GitHub Pages (landing page + dashboard)
├── docs/ # Architecture diagrams (D2 format)
├── kafka_to_postgres.py # Kafka consumer → TimescaleDB
├── live_from_kafka.py # Real-time Kafka producer
├── Dockerfile.kafka # Kafka producer/consumer image
├── Dockerfile.airflow # Custom Airflow image
├── docker-compose.yml # 7-service orchestration
├── ruff.toml # Linter and formatter config
├── Makefile # Common commands (up, down, test, lint)
├── .env.example # Environment variable template
└── requirements.txt # Pinned Python dependencies
| Layer | Technology |
|---|---|
| Streaming | Apache Kafka (Confluent) with Zookeeper |
| Orchestration | Apache Airflow 2.7 |
| Database | TimescaleDB (PostgreSQL 14 + time-series extensions) |
| Containerization | Docker Compose (7 services) |
| CI/CD | GitHub Actions (ruff lint + pytest) |
| Frontend | GitHub Pages + Chart.js |
| Language | Python 3.12 locally; the Airflow image is python 3.11 |
Contributions are welcome! See CONTRIBUTING.md for setup instructions and guidelines.
This project is licensed under the MIT License — see the LICENSE file for details.