How kedarg63/hedgefund works
quantdata is a primary-source research-data pipeline and point-in-time research terminal for Indian and US equities, commodities, and macro data. Its central workflow collects exchange, regulator, ministry, and broker data; archives original bytes; parses them into a knowledge-date-stamped parquet warehouse; and exposes research views through a Streamlit dashboard and authenticated API/React terminal. Analytics and quality monitoring support derived signals and trust checks. The README documents additional connector and analytics capabilities not represented by sampled implementation excerpts; those are shown at subsystem level, with uncertain internal wiring omitted.
Main entry points: storage.py Parquet warehouse · run_daily.py Daily jobs · connectors India connectors · stream.py Live stream
Components 22
Data collection
- India sources
- US sources
- Global and macro sources
- Daily jobs
run_daily.py - India connectors
connectors - US filing connectors
connectors - Macro and commodity connectors
connectors
Archive and warehouse
- Immutable raw archive
core/storage.py - Parquet warehouse
core/storage.py - DuckDB views
core/storage.py
Research and monitoring
- Signal analytics
analytics - Data quality
core/quality.py - Failure alerts
core/alerting.py
Research interfaces
- Streamlit dashboard
dashboard/app.py - Terminal API
api/main.py - Data routes
api/routers - React terminal
web/src/App.tsx - Research panels
web/src/panels
Live market data
- Live stream
connectors/stream.py - Tick cache
core/redis_client.py
Other components
- Researcher
- Broker feed
Connections 27
- Daily jobs → India connectors: runs jobs
- Daily jobs → US filing connectors: runs jobs
- Daily jobs → Macro and commodity connectors: runs jobs
- India sources → India connectors: supplies data
- US sources → US filing connectors: supplies filings
- Global and macro sources → Macro and commodity connectors: supplies data
- India connectors → Immutable raw archive: archives bytes
- US filing connectors → Immutable raw archive: archives bytes
- Macro and commodity connectors → Immutable raw archive: archives bytes
- India connectors → Parquet warehouse: writes parsed data
- US filing connectors → Parquet warehouse: writes parsed data
- Macro and commodity connectors → Parquet warehouse: writes parsed data
- Parquet warehouse → DuckDB views: backs views
- Signal analytics → Parquet warehouse: writes derived data
- Streamlit dashboard → DuckDB views: queries views
- Streamlit dashboard → Data quality: checks freshness
- Daily jobs → Failure alerts: reports failures
- Terminal API → Data routes: registers routes
- Data routes → DuckDB views: queries as-of data
- React terminal → Terminal API: requests data
- React terminal → Research panels: hosts panels
- Researcher → Streamlit dashboard: explores research
- Researcher → React terminal: uses terminal
- Broker feed → Live stream: streams ticks
- Live stream → Tick cache: publishes ticks
- Live stream → Immutable raw archive: archives stream
- Tick cache → Terminal API: serves live data