Time Series Insights - Univariate

This analysis helps in understanding the characteristics of a single time series feature, including autocorrelation, stationarity, and lag relationships.

Graph Type:

• ACF (Autocorrelation Function): Measures correlation with past values over different lags.
• PACF (Partial Autocorrelation Function): Shows direct correlation between a time series and its past values, excluding intermediate lags.
• Lag Analysis: Identifies the optimal lag order to understand past influence on future values.

Feature: Select a specific time-dependent variable for analysis, such as "temperature" in a climate dataset.

Order Differencing: Used to remove trends and make the series stationary.

• First-order differencing (d=1): Removes linear trends.
• Second-order differencing (d=2): Removes quadratic trends.

Custom Order: Manually set the differencing order for better stationarity, useful when automatic differencing is insufficient.

This analysis aids in selecting the right forecasting model, such as ARIMA, by identifying significant patterns and trends.