Forecast sales and find what drives them
Two analyst tools in one, with plain-English results.
Forecast
- Groups your data by day, week, month or quarter (missing periods count as zero).
- Tries a straight-line trend, Holt's trend smoothing and Holt-Winters (trend + seasonality), and picks the one that fits your history best (lowest AIC).
- Shows the forecast with a 95% range, the method's average error on past data, and a downloadable table.
What drives it? (regression)
- Multiple linear regression (ordinary least squares) with numeric drivers and categories (for example region).
- Coefficients, standard errors, t-values, p-values, 95% ranges, R², adjusted R² and the F-test, all explained in plain English.
- A "try a prediction" box to test what-if scenarios.
The maths was checked against the Python statsmodels and NumPy libraries.