MTFCK! (NSE MTF Analytics Dashboard)

A robust and interactive dashboard for analyzing NSE Margin Trading Facility (MTF) data. Gain insights into top stocks by amount financed, exposure percentages, returns, and interactive market trends.

🌟 Key Features

🚀 Macro level analysis

See the big picture by comparing the Total Outstanding Market Trend with the NIFTY TOTAL MARKET movement over time, allowing you to gauge the overall market leverage environment.

Provides four-level industry filter with granular multi-select industries. Search among 200 industries over any four levels. Choose one or more basic industries. eg. search Discretionary and see all Discretionary industries and basic industries. Hover over a basic industry to understand its hierarchy

View and sort market leaders by Amount Financed, % Change, and Exposure %. Easily track newly added MTF stocks and deep-dive into the data using the smart industry hierarchy filters.

Exposure provided as a % of free float market cap to help understand the actual levered exposure liquidation risk

📈 Micro level analysis

Analyze individual stock performance by overlapping the Amount Financed Trend directly with the Closing Price to identify correlations and market leverage points. Leverages a robust database that keeps track of symbol changes and merges data, providing continuous, seamless data. No fragmentation

Rich metadata provided for at-a-glance preliminary decisions in all screens(returns for key time periods, exposure, free float market cap etc)

Date range filters respected by all functions

Functionality-rich components:

• graphs have autosize, zoom, selection of each charts,
• tables have export functions, sortability etc

⚡ Under the Hood

• Automated Data Pipeline: Daily updates via GitHub Actions in the MTFDB repository.
• Parquet Data Lake: Uses hyper-compressed Parquet files for storage, reducing 200MB+ of raw CSV data into ~2.5MB.
• One-Click Sync: Synchronize your local dashboard with the latest cloud data directly from the UI.
• Performance Optimized: Uses DuckDB in-memory for lightning-fast analytical queries directly on Parquet files.
• Smart Industry Filtering: Searchable industry hierarchy for deep sector analysis.

Usage

1. Install dependencies

   uv sync

2. Run the dashboard

   uv run streamlit run app.py

3. Controls

   • Sync Database: Click "Sync Database from Cloud" in the sidebar to download the latest data.
   • Analysis: Select date range, industry, and analysis type, then click "Run Analysis".
   • Trends: Enter a symbol and click "Show Amount Financed Trend" to see historical MTF data overlapped with price.
   • Total outstanding: Click "Show Total Outstanding Trend" to see the market-wide leverage trend overlapped with the NIFTY Total Market index.

Project Structure

• app.py - Streamlit dashboard UI and cloud sync logic.
• src/mtfck/mtfck.py - Core analytical logic and Just-In-Time (JIT) industry enrichment.
• src/mtfck/ingestion.py - Data pipeline for parsing NSE reports and managing Parquet exports.
• src/mtfck/db.py - Database connection manager (In-memory DuckDB mapped to Parquet).
• mtf_data/ - Local cache for .parquet data files (ignored by Git).

Tech Stack

• UI: Streamlit
• Analytics: DuckDB (In-process SQL engine)
• Storage: Apache Parquet (ZSTD Compressed)
• Data Pipeline: Python + GitHub Actions
• Data Source: NSE India (National Stock Exchange)

License

MIT License