Demand Forecasting Simulator

Businesses live or die by how well they predict what customers will buy next. This simulator generates a rolling series of weekly demand — a baseline that drifts with a trend, rises and falls with a seasonal cycle, and jitters with random volatility — and forecasts it live using simple exponential smoothing, the same technique described in the accompanying article. Watch the green actual-demand bars and amber forecast bars race each other week by week, and tune the smoothing constant, seasonality and volatility to see how each reshapes forecast accuracy (MAPE) and how quickly the model reacts to real shifts versus noise.