Skip to content
LeadAfrik
Advanced · Self-paced2026 Edition

Time Series for Finance

The time-series machinery a quant uses every day — ARMA, GARCH, cointegration, VAR, and state-space models — taught with the financial intuition that makes the algebra stick. Eleven modules from stationarity to Kalman filtering, building toward a volatility model on NSE returns and a pairs-trading cointegration test you can defend.

11

Modules

~11h 15m

Reading time

Advanced

Level

Self-paced

Format

§

Syllabus

  1. 01

    Stochastic processes and stationarity

    What a time-series actually is. Strict vs weak stationarity, ergodicity, autocovariance, the autocorrelation function (ACF).

    ~55 minModule 01
    →
  2. 02

    White noise, random walks, and why returns ≠ prices

    Price series are typically non-stationary; returns are typically stationary. The unit root that separates them.

    ~50 minModule 02
    →
  3. 03

    AR, MA, and ARMA models

    Autoregressive, moving-average, and ARMA — the building blocks. Stationarity and invertibility conditions, the lag operator.

    ~60 minModule 03
    →
  4. 04

    Identification, ACF/PACF, AIC/BIC

    Box-Jenkins identification, the partial autocorrelation function, information criteria, and the residual diagnostics you actually run.

    ~55 minModule 04
    →
  5. 05

    Forecasting and the Wold decomposition

    Point forecasts, forecast errors, prediction intervals. Wold's theorem and why every stationary series is essentially an MA(∞).

    ~55 minModule 05
    →
  6. 06

    Unit roots, ARIMA, and cointegration

    Dickey-Fuller and KPSS tests, differencing, Engle-Granger and Johansen cointegration — the algebra behind pairs trading.

    ~65 minModule 06
    →
  7. 07

    Volatility — ARCH, GARCH, EGARCH

    Why financial volatility clusters. ARCH/GARCH(1,1), asymmetric models, the leverage effect, fitting a GARCH on NSE returns.

    ~65 minModule 07
    →
  8. 08

    Multivariate time series — VAR and VECM

    Vector autoregressions, Granger causality, impulse-response functions, VECM for cointegrated systems.

    ~60 minModule 08
    →
  9. 09

    State-space models and the Kalman filter

    Hidden states, observation equations. The Kalman filter as recursive Bayesian estimation. Smoothing, missing data, time-varying parameters.

    ~65 minModule 09
    →
  10. 10

    Dynamic factor models and PCA on returns

    Reducing a high-dimensional return panel to a handful of factors. Yield-curve PCA, equity factor extraction, the Nelson-Siegel parametrisation.

    ~55 minModule 10
    →
  11. 11

    Three real time-series cases

    Build a GARCH vol model on NSE-20 returns; test cointegration between Safaricom and the NSE; fit a Kalman filter to a noisy commodity series.

    ~90 minModule 11
    →

How to use this course

Start with module 01 if the material is new; skip ahead if you have prior exposure. Each module is self-contained but the arc is sequential — the projects in the final module assume the toolkit from modules 1-11. Every module ends with key takeaways and a curated further-reading list with primary sources.