Home▸Data Science▸ACF/PACF Order Identification: Reading a Time Series' Autocorrelation Signature

ACF/PACF Order Identification: Reading a Time Series' Autocorrelation Signature

Interactive Box-Jenkins simulator: generate AR, MA, or random-walk time series and watch the autocorrelation (ACF) and partial autocorrelation (PACF) bar charts, live, in 3D — with the classic cutoff/decay signature you use to pick an ARIMA(p,d,q) order.

Data Science3DAdvanced60 FPS

Choosing an ARIMA(p,d,q) order isn't guesswork — it's reading two bar charts. This simulator generates a synthetic time series from a white-noise, AR(1), AR(2), MA(1), or non-stationary random-walk process, computes its autocorrelation function (ACF) and partial autocorrelation function (PACF) via the Durbin–Levinson recursion, and renders both as live 3D bar charts against the ±1.96/√N significance band. Drag the process coefficients to watch the classic Box-Jenkins signatures appear in real time — an AR process's PACF cutting off sharply while its ACF decays, an MA process doing the reverse, and a random walk's ACF refusing to decay until you flip on first-order differencing.

⚙ Under the hood

Generate AR, MA, or random-walk time series and watch their autocorrelation (ACF) and partial autocorrelation (PACF) bar charts update live in 3D — the classic cutoff/decay signature analysts use to pick an ARIMA(p,d,q) order.

time seriesARIMAautocorrelationPACFBox-Jenkinsstationarity

3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install