Sensor data trends are commonly used by engineers to get an overview of process dynamics and history. Reading and description of these trends is understood as a process of segmentation and categorization of time series in a sequence of symbols. Based on this procedure we present a method for extraction and analysis of trends in process data corresponding to the operator's view of signal processing. The idea of interaction and algorithmic implementation are presented. Finally the application to experimental data of a distillation column shows the capability to provide assistance in exploration of historical data for analysis, diagnosis and decision-making in operation of chemical processes.