Design of Exploratory Data Analysis Tools for Thermodynamic Concepts in the Steam Power Plant Process Using the Streamlit Framework

Thermodynamics is often regarded as an abstract and challenging subject by many students, largely due to the complexity of its core principles. Concepts such as entropy, enthalpy, and energy transformation are typically introduced through detailed mathematical equations, which are not always accompanied by adequate visual explanations. This lack of visual support can make it difficult for students to fully grasp the underlying phenomena. To address this issue, the present study focuses on the development of a visualization tool that leverages experimental and simulation data to enhance the learning experience in thermodynamics. The tool is built using operational data collected from a mini coal-fired power plant trainer, which serves as a hands-on educational platform for simulating thermodynamic processes. The application is developed in the Python programming language and deployed using the Streamlit web framework. Several features are integrated into the tool, including automated data reporting, correlation analysis visualized through heatmaps, dynamic visualization of energy flow using Sankey diagrams, and the implementation of three machine learning models to predict electrical output. To evaluate the effectiveness of the tool, experts in the field of thermodynamics assessed it using a Likert scale based on four main criteria. The evaluation results indicate strong performance in all categories. The accuracy of information received an average score of 4.0, visual quality was rated at 4.25, alignment with educational objectives scored 3.75, and ease of use was rated highest at 4.5. All scores fall within the “Good” category, with some approaching “Very Good.” These findings suggest that a data-driven visualization tool, when based on realistic and practical datasets, holds significant potential for improving the clarity and interactivity of thermodynamics instruction.