Use this tool to see how strongly each observation of a time series is related to its own past values. Paste one value per line (or date,value rows), choose the maximum lag and any differencing, and read the autocorrelations, 95% limits and Ljung–Box tests. Change the colours, text and line sizes, then sign in free to download the figure and an APA 7 Word report.
About this tool
The autocorrelation function (ACF) correlates a series with lagged copies of itself. Spikes outside the 95% limits show lags at which the series is still related to its past. A slow decay suggests a trend or non-stationarity, a repeating pattern at lag 12 suggests monthly seasonality, and a sharp cut-off after lag q points to a moving-average (MA) model of order q.
This tool computes the sample autocorrelations, draws the correlogram with approximate 95% confidence limits (±1.96/√n), runs the Ljung–Box portmanteau test for white noise and gives plain-language guidance on stationarity and MA order. First differences and seasonal differences can be applied before the plot.
When to use it
Use an ACF plot before fitting ARIMA models, when checking residuals for leftover autocorrelation, when looking for seasonality in monthly or weekly counts, and when deciding whether differencing is needed. It suits economics, epidemiology, ecology and agricultural monitoring data.
What this tool reports
Autocorrelations at lags 1 to the maximum lag with 95% limits.
Ljung–Box Q statistic, degrees of freedom and p value.
Stationarity and MA-order guidance in plain language.
Optional first or seasonal differencing before the plot.
Editable APA 7 results text and tables, plus a figure you can restyle.
Assumptions to check
Observations are equally spaced in time and in time order, with no missing values.
At least 30 observations are recommended for stable estimates at higher lags.
The limits are approximate and assume a white-noise null; one spike in 20 lags can occur by chance.
How to report it in APA 7
Report the series, the number of observations, the lags examined and the Ljung–Box result, for example Q(24) = 41.3, p = .015, then describe which lags exceed the limits and what that implies for stationarity or model order. The tool drafts this text and an APA 7 table that you can edit before downloading.
Downloads
Analyses and figures are free to view. Create a free account or log in to download the figure as a picture (PNG or SVG) or the combined APA 7 results and figure as a Word (.docx) file. Colours, text colour, text size and line size can be changed before you download.
Frequently asked questions
Is the ACF plot generator free?
Yes. You can analyse your series and view the figure and results without an account. A free account is needed only to download the figure, tables or Word report.
How many lags should I plot?
A common rule is about 10 lags, or one to two seasonal cycles (for example 24 for monthly data), and no more than a quarter of the number of observations.
What is the difference between ACF and PACF?
The ACF shows the total correlation at each lag, including indirect effects through shorter lags. The PACF removes those indirect effects and is mainly used to choose the autoregressive order.
Is my data uploaded?
No. The analysis runs in your browser and your data stays on your device.