In modern industrial environments, sensors play a crucial role for automation by continuously analyzing large volumes of time series data vital for process optimization. However, analyzing this data in isolation poses significant challenges, particularly in time series analysis, due to the influence of external contextual factors that are not always directly observable. Integrating these is essential for time series analysis. While, data fusion is a technique that aims at integrating or blending data with different modalities for time series analysis, such as images or videos, contextual factors may not always be heterogeneous in modality, but rather heterogeneous in time dimension, which makes its integration challenging. Therefore, we identified four different types of time dimensions that often appear in industrial environments, namely constant, time series, events, and intervals, and we aim at introducing the foundation towards a systematic approach for integrating contextual factors with heterogeneous time dimensions. This enables the transformation of data with heterogeneous time dimensions into a format that can be effectively processed by traditional machine learning models for time series analysis.