Time Series & Forecasting
Time series work is where modelling assumptions meet time, operations and data collection. This hub emphasizes forecasting decisions, validation, seasonality, anomalies and the practical traps that make backtests too optimistic.
Start here
- Forecasting Baselines That Are Hard to Beat
- Feature Engineering for Time Series Without Leaking the Future
- Missing Data and Irregular Sampling in Time Series
- State Space Models and the Kalman Filter
Forecasting practice
Baselines, probabilistic forecasts, hierarchical reconciliation, intermittent demand and model evaluation.
State Space Models and the Kalman Filter
The Kalman filter is usually introduced as a tracking algorithm for spacecraft. It is more useful understood as the general engine for estimating hidden state from noisy observation.
Global vs Local Models in Time Series Forecasting
The traditional approach fits one model per series. Modern practice often fits a single model across thousands of them, and usually wins.
Anomaly Detection in Time Series
Outlier detection asks whether a value is unusual. Time series anomaly detection asks whether it is unusual now, which is a different and harder question.
Probabilistic Forecasting: Beyond the Point Estimate
A point forecast answers the wrong question. Most decisions depend on how bad things could plausibly get, which is a statement about the whole distribution.
Missing Data and Irregular Sampling in Time Series
A missing row in a table is a nuisance. A missing interval in a time series changes the meaning of every lag, window and seasonal index computed from it.
Hierarchical Forecasting: Making Forecasts Add Up
Forecast every store separately and the total will not match the forecast you made for the company. Reconciliation is how you make a hierarchy of forecasts coherent.
Feature Engineering for Time Series Without Leaking the Future
Turning a time series into a tabular problem unlocks powerful models and introduces a specific failure: features that quietly contain information from the future.
Multiple Seasonality: MSTL, TBATS, and Fourier Terms
Hourly and daily data rarely has one season. Electricity demand cycles daily, weekly and annually at the same time, and a single seasonal period cannot represent that.
Forecasting Baselines That Are Hard to Beat
An RMSE of 4.2 means nothing on its own. Without a baseline you cannot tell whether a model is skilful or merely arithmetic.
Intermittent Demand Forecasting: Croston’s Method and Its Successors
Spare parts and slow-moving stock produce series that are mostly zeros. Standard forecasters quietly fail on them; Croston’s method and its successors are built for exactly this shape of data.
Evaluating Time Series Forecasting Models: Metrics and Best Practices
Effective model evaluation is essential for reliable time series forecasting. Learn the most important metrics, validation methods, and strategies for interpreting and improving forecasts.
Multivariate Time Series Forecasting: VAR and VECM Models Explained
A practical guide to VAR and VECM for multivariate time series forecasting, including math, assumptions, cointegration testing, and Python code.
Diagnostics and data quality
Non-stationarity, missing intervals, anomaly detection, feature leakage and seasonal structure.
State Space Models and the Kalman Filter
The Kalman filter is usually introduced as a tracking algorithm for spacecraft. It is more useful understood as the general engine for estimating hidden state from noisy observation.
Global vs Local Models in Time Series Forecasting
The traditional approach fits one model per series. Modern practice often fits a single model across thousands of them, and usually wins.
Anomaly Detection in Time Series
Outlier detection asks whether a value is unusual. Time series anomaly detection asks whether it is unusual now, which is a different and harder question.
Probabilistic Forecasting: Beyond the Point Estimate
A point forecast answers the wrong question. Most decisions depend on how bad things could plausibly get, which is a statement about the whole distribution.
Missing Data and Irregular Sampling in Time Series
A missing row in a table is a nuisance. A missing interval in a time series changes the meaning of every lag, window and seasonal index computed from it.
Hierarchical Forecasting: Making Forecasts Add Up
Forecast every store separately and the total will not match the forecast you made for the company. Reconciliation is how you make a hierarchy of forecasts coherent.
Feature Engineering for Time Series Without Leaking the Future
Turning a time series into a tabular problem unlocks powerful models and introduces a specific failure: features that quietly contain information from the future.
Multiple Seasonality: MSTL, TBATS, and Fourier Terms
Hourly and daily data rarely has one season. Electricity demand cycles daily, weekly and annually at the same time, and a single seasonal period cannot represent that.
Forecasting Baselines That Are Hard to Beat
An RMSE of 4.2 means nothing on its own. Without a baseline you cannot tell whether a model is skilful or merely arithmetic.
Intermittent Demand Forecasting: Croston’s Method and Its Successors
Spare parts and slow-moving stock produce series that are mostly zeros. Standard forecasters quietly fail on them; Croston’s method and its successors are built for exactly this shape of data.
Fourier Analysis for Data Science: From Signals to Features
Fourier analysis is more than a signal-processing trick. It is a way to ask which cycles, rhythms, and scales explain variation in data.
Anomaly Detection in Sensor Streams
Sensor anomaly detection works best when statistical signals, domain constraints, and alert workflows are designed together.