This paper discusses the characteristics of process industry data critical for developing data-driven soft sensors. Soft sensors are predictive models that provide information similar to traditional sensors. They are used for process monitoring, fault detection, and control. The paper highlights the importance of data-driven soft sensors due to their growing popularity and potential. It discusses the challenges in developing these sensors, including data issues like missing values, outliers, and drifting data. The paper also covers the most common modelling techniques used in soft sensor development, such as Principal Component Analysis (PCA), Partial Least Squares (PLS), Artificial Neural Networks (ANN), and Support Vector Machines (SVM). It discusses the application of soft sensors in various process industries, including chemical, bioprocess, and steel industries. The paper also addresses the maintenance and adaptation of soft sensors, emphasizing the need for regular updates to maintain performance. The paper concludes with a discussion of future research directions in soft sensor development.This paper discusses the characteristics of process industry data critical for developing data-driven soft sensors. Soft sensors are predictive models that provide information similar to traditional sensors. They are used for process monitoring, fault detection, and control. The paper highlights the importance of data-driven soft sensors due to their growing popularity and potential. It discusses the challenges in developing these sensors, including data issues like missing values, outliers, and drifting data. The paper also covers the most common modelling techniques used in soft sensor development, such as Principal Component Analysis (PCA), Partial Least Squares (PLS), Artificial Neural Networks (ANN), and Support Vector Machines (SVM). It discusses the application of soft sensors in various process industries, including chemical, bioprocess, and steel industries. The paper also addresses the maintenance and adaptation of soft sensors, emphasizing the need for regular updates to maintain performance. The paper concludes with a discussion of future research directions in soft sensor development.