Real data screams with noise—seasonal spikes, random wobbles, cycles. Your job isn't to listen to all of it. It's to subtract the predictable patterns (seasonality, cycles) and smooth the chaos so the underlying trend emerges clean. Drag the slider to increase the smoothing window and watch noise collapse into clarity.
Raw Time Series + Smoothing
Window: 1 period
Drag slider to change moving average window size
Noise dominates • No trend visible
Smoothing Window
1612
Window Size
1
Noise Reduction
0%
Time series data has four components: trend (the long-term direction), seasonality (predictable repeating patterns like summer spikes), cycles (longer waves tied to economic or environmental forces), and irregular variation (random noise). None of these sit still—they overlap, obscure each other. To forecast anything, you must separate them. Moving averages smooth out the irregular bumps; the longer the window, the more noise disappears, revealing the trend beneath.
Know This
Moving mean/median smoothing strips irregular noise to reveal the trend; seasonal indices measure each season's typical effect so you can deseasonalise the data and forecast the clean underlying pattern.