Quant

research models
build model (math)
build model (python)
get historical data
test model
use free stock price api (yahoo finance)
test model
report

  1. Vectorized Mathematics

  2. NumPy Arrays: Multi-dimensional arrays, slicing, boolean masking, memory layouts (C vs Fortran contiguous).

  3. Array Math: Element-wise operations, broadcasting rules, matrix multiplication (@).

  4. Linear Algebra: Matrix inversion, eigenvalues/eigenvectors, Singular Value Decomposition (numpy.linalg).

  5. Financial Time-Series

  6. Pandas Structs: Series and DataFrames with DatetimeIndex.

  7. Alignment: merge_asof for irregular tick alignment, handling look-ahead bias via forward-filling (ffill).
  8. Window Operations: rolling(), expanding(), and ewm() for moving averages, volatility, and tracking signals.

  9. Resampling: Converting raw trade tick streams into open-high-low-close (OHLC) bars (resample()).

  10. Statistical Modeling & Backtesting

  11. Regression: Ordinary Least Squares (OLS), rolling regressions, asset beta calculations (statsmodels).

  12. Machine Learning: Cross-validation for time-series, feature scaling, classification metrics (scikit-learn).

  13. Simulations: Monte Carlo path generation, vector-based backtesting engines (vectorbt).

  14. Performance Optimization

  15. Vectorization: Elimination of standard Python for loops in favor of contiguous array operations.

  16. JIT Compilation: Using @njit decorators to compile mathematical loops into machine code (numba).

  17. Parallelism: Circumventing the Global Interpreter Lock (GIL) via multiprocessing for concurrent backtests.

  18. Technical Tooling

  19. Data Sources: API interactions, CSV streaming, parquet file serialization for rapid I/O.

  20. Core Libraries: numpy, pandas, scipy.optimize, statsmodels, scikit-learn, numba.