Demand forecasting studio

Meridian Foods Forecast Studio

Python-based statistical forecasting with a simple static dashboard. Sample data stays local, forecasts are prebuilt into JSON, and the front end remains lightweight and easy to host.

Engine pandas · statsmodels
Models EMA · Holt-Winters · AR(3)
Delivery Static HTML dashboard
Loading forecast payload…

Demand history and forecast

Actual, outlier-adjusted history, EMA smoothing, and forward forecast with confidence bands.

Series

Model comparison

Compare the next horizon across the three component models and the ensemble.

Weighting and fit

Static ensemble weights plus holdout RMSE from the Python modeling step.

Detected outliers

Rolling z-score anomalies replaced with the local rolling median before modeling.

Date Type Raw Adjusted Z-score