Use this tool when you have a series of values recorded in time order, such as monthly clinic visits, weekly sales or yearly rainfall, and you want a forecast with an honest uncertainty band. Upload an Excel, CSV or SPSS file (or paste your values), choose the seasonal period and forecast horizon, and click Analyze data to get a Holt–Winters forecast, accuracy on a holdout sample and residual checks, written up in APA 7 format.
About forecast plots and Holt–Winters forecasting
A forecast plot shows a time series together with its fitted values, the predicted future values and a prediction interval that widens as the forecast looks further ahead. This tool uses Holt–Winters exponential smoothing, which tracks the level, trend and (optionally) a repeating seasonal pattern and gives more weight to recent observations. The smoothing parameters are chosen from your data, and the tool falls back to Holt’s linear trend when the series is too short for the seasonal period.
When to use it
Use it for a single numeric series observed at regular intervals, with at least a few full seasonal cycles when the data are seasonal. It suits public health, education, business, ecology and agriculture reports where a short-term projection with an interval is needed. It is not suited to series with irregular gaps, missing values or sudden structural breaks.
What this tool reports
A forecast plot with the observed series, fitted values and a 95% prediction interval.
Holdout accuracy: RMSE, MAE and MAPE on the last part of the series.
Residuals over time and the residual autocorrelation plot (ACF) with confidence limits.
The Ljung–Box test for remaining autocorrelation in the residuals.
Seasonal period and forecast horizon settings, with a holdout check on the last observations.
Editable APA7 results text, APA7-style tables and the figure, ready to download.
Assumptions to check
Observations are equally spaced and in time order, with no missing values.
The seasonal period is set correctly (for example 12 for monthly data with a yearly cycle).
Residuals should look like random noise; check the residual plot and Ljung–Box result before trusting the interval.
How to report it in APA7
Describe the series and its length, name the model (Holt–Winters, additive or multiplicative), report the smoothing parameters, the holdout accuracy (RMSE, MAE, MAPE) and the Ljung–Box result, then state the forecast for the horizon with its 95% prediction interval. The tool writes this paragraph for you and you can edit it before downloading.
Frequently asked questions
Is the forecast plot generator free?
Yes. You can run the analysis, see the figure and read the APA7 results without an account. Create a free account or log in only when you want to download the results.
How many observations do I need?
For seasonal data use at least two full seasonal cycles, and preferably three or more. For non-seasonal data about 20 or more observations give more stable estimates.
What does the 95% prediction interval mean?
It is the range in which a future observation is expected to fall about 95% of the time if the model assumptions hold. It is wider than a confidence interval for the mean because it includes the random variation of a single new value.
Is my data uploaded to a server?
The analysis runs in your browser, so your data stays on your device while you work with the tool.