Abstract
In the ever-evolving educational milieu, the integration of innovative teaching methodologies is increasingly crucial to meet the changing needs of modern learners. This research meticulously explores the application of open teaching practices in the fields of building environment and energy application engineering. Through an in-depth examination of multi-regression data pertaining to various environmental factors, this study reveals significant correlations and patterns that are relevant to both educators and environmental specialists. Emphasis is placed on the student-centric ethos of this approach, combining the dual concepts of environmental science and pedagogical progression. The relationship between environmental variables, such as PM2.5, PM10, temperature, and humidity, and the air quality index (AQI) is rigorously analyzed. Such analysis underscores the educational improvements brought about by open teaching strategies. The presented findings not only offer nuanced insights into how the aforementioned variables influence air quality but also highlight the benefits and potential of open teaching methodologies in creating a more interactive and enlightening academic environment.
Introduction
As global urbanization accelerates, the challenge of creating healthy, energy-efficient indoor spaces has become paramount. This requires a sophisticated understanding of the complex, interacting variables that dictate the Air Quality Index (AQI). However, traditional engineering education often relies on static, theoretical models that fail to prepare students for these real-world complexities.
This paper addresses this gap by exploring the intersection of environmental engineering and modern pedagogy. By utilizing "open teaching"—an approach emphasizing accessibility, open-source resources, and community collaboration—the authors demonstrate how real-world environmental data analysis can serve as a powerful vehicle for student engagement and technical skill development in Building Environment and Energy Application Engineering.
Methodology and Data Collection
The research utilizes a two-pronged methodology involving environmental modeling and the implementation of open teaching practices. Data was collected between October 2016 and December 2017 using precision sensors to monitor temperature, humidity, particulate matter (PM2.5, PM10), $CO_2$, and illumination.
Environmental Modeling
The study moved beyond simple linear models to analyze the temporal dynamics of indoor and outdoor parameters. While the authors explored nonlinear regression models—such as $E = \alpha \times T^n + \beta \times e^H + \gamma \times \ln(A)$—the core of the analysis focused on robust multi-regression equations tailored to specific seasonal timeframes. Data preprocessing included moving average smoothing to reduce noise and Min-Max normalization for scaling.
Open Teaching Implementation
The curriculum was redesigned to incorporate Open Educational Resources (OERs). Students utilized collaborative platforms like Moodle and Edmodo for discourse and GitHub for version control and collaborative coding. This transparency fostered a culture of collective problem-solving, mirroring the operations of modern engineering firms.
Key Results and Insights
The multi-regression models proved highly robust across different timeframes, achieving high predictive accuracy. The authors reported R-squared values ranging from 0.88 to 0.92, with a Mean Absolute Error (MAE) of 2.34 and a Root Mean Square Error (RMSE) of 3.21.
Significant Determinants
Analysis revealed that PM2.5 and temperature were consistently the most influential factors on air quality. A key "aha moment" was the discovery of synergistic interactions; for instance, the relationship between temperature and humidity has a non-linear and significant impact on measured air quality. Furthermore, the weight of certain predictors was found to be seasonally sensitive, with illumination having a lower impact during winter months.
Pedagogical Impact
Qualitatively, the hands-on, open-resource strategy led to significantly higher levels of student involvement. By transforming students from passive consumers of formulas into active researchers analyzing live data, the "open teaching" approach improved both practical analytical skills and conceptual grasp.
Conclusion and Future Work
While the multi-regression models are effective, they represent a starting point for more advanced environmental analytics. Future research should expand the variable set to include volatile organic compounds (VOCs) and nitrogen dioxide ($NO_2$). Additionally, the authors suggest transitioning from regression models toward machine learning (ML) algorithms to further enhance predictive accuracy. Ultimately, these findings provide a roadmap for both educators seeking to modernize engineering curricula and policy architects aiming to design healthier urban environments.
Summary updated