Time Series Forecasting

Model sequential temporal changes. Smooth historical noise using moving averages or exponential factors, and fit linear trend regressions to forecast future intervals.

Data Point Manipulation & Smoothed Projections

Drag any of the solid blue historical data points vertically to modify values. The smoothed fit line (emerald) and forecasted projection (dotted orange) will update in real-time.

1. Data Shape Preset

2. Smoothing Algorithm

Method:
Forecast Horizon:
SMA Period (k months): 3

3. Diagnostics Summary

Mean Absolute Error (MAE): MAE = 4.31
Linear Trend Line: y = 1.25x + 35.4

Smoothing & Forecasting Trace Walkthrough

Trace how historical noise smoothing algorithms are calculated and how slope parameters are fitted to project future values.

Trace Calculations Steps

Historical vs. Smoothed Data Table


Time Series Quiz

Test your conceptual knowledge of trends, cyclical seasonality, smoothing variables, and phase lags.

Question 1 of 5 Score: 0/0

Quiz question loading...

Correct Answer!

Explanation text...

Time Series Cheat Sheet


  • Trend: The long-term upward or downward direction of data (modeled as $y = mx + c$).
  • Seasonality: Repeating cycles that occur within a calendar year or fixed interval (e.g. summer sales spikes).
  • Smoothing: Moving averages filter out short-term random noise to reveal trends.
  • Phase Lag: Smoothed lines always lag behind original peaks because they incorporate past historical data points.