Builds a stock universe from 15 NSE benchmark and sectoral index constituent lists, then classifies every individual stock and every index (via its member stocks) into a momentum zone — Highly or Moderately Bullish or Bearish.
Uses a Clenow-style annualized-slope momentum score (slope of log-price × R²) blended with 5/20/50-day ROC velocity.
Fixed in review
- The source notebook had grown to four "Features" (Alpha–Delta). Features Gamma and Delta were discarded entirely because they fabricated data — np.random-based fake "bullish probabilities" and Poisson-sampled fake pattern counts, dressed up as a "COMPLETE WORKING VERSION" screener without ever touching real price data.
- Of the two real features kept: a manual 30-CSV Colab upload replaced with a programmatic NSE-archive fetch, and blocking input() prompts removed.
- Eliminated a duplicated, harder-to-verify index-ticker mapping (including one ticker that looked mislabeled) by deriving index-level zones from the same single stock-level fetch.
- Fixed a rank/remaining-count bug in both 4-box visualizations.
- Fixed an arbitrary alphabetical .head(100) sample that silently meant "companies A through ~M" rather than a representative sample.
- Dropped two non-functional leftover cells.
Clenow momentumClassificationIndex constituentsNSE