Introduction
As Data Science and Engineering (DSE) becomes increasingly critical across all technical disciplines, undergraduate engineering programs face a significant challenge: how to integrate these essential skills without overburdening an already dense curriculum. Researchers at Auburn University are addressing this by developing DEEP (Data-enabled Engineering Project) modules. These interactive, web-based tutorials are designed to bring real-world, "messy" data into existing engineering courses, bridging the gap between theoretical knowledge and practical application.
The Challenge: Moving Beyond "Canned" Data
Traditional engineering education often relies on "canned" datasets—perfectly cleaned, curated, and paired with specific algorithmic instructions. While useful for teaching basic mechanics, this approach fails to prepare students for the reality of professional engineering, where data is often noisy, incomplete, or ill-posed. Furthermore, the rigid structure of engineering degrees makes adding dedicated DSE courses administratively difficult. The DEEP framework seeks to solve this by providing authentic research experiences that can be seamlessly integrated into core engineering subjects like Thermodynamics or Circuits.
The DEEP Framework and Methodology
The project utilizes an IDLE (Interactive Development and Learning Environment) to deliver modular content. The pedagogical approach is grounded in two key educational theories:
- Experiential Learning Theory (ELT): Following Kolb’s learning cycle, students progress through concrete experience, reflective observation, abstract conceptualization, and active experimentation.
- Course-based Undergraduate Research Experience (CURE): The modules function as "short CUREs," offering the benefits of research—such as discovery and iteration—within a manageable, modular format.
To ensure accessibility and broad adoption, the modules are developed across multiple platforms, including Jupyter Notebooks (Python), MATLAB Live Scripts, and SAS. By leveraging cloud-based environments like Google Colab, the team eliminates the need for complex local software installations, allowing students to focus on the data rather than the environment.
Data Sources and Real-World Application
A cornerstone of the DEEP modules is the use of diverse, high-fidelity datasets to simulate real engineering challenges:
- IoT Testbeds: Data generated from two custom-built Internet-of-Things enabled laboratory testbeds.
- Engineering Benchmarks: Standard datasets from MathWorks covering chemical reactions, automotive measurements, and medical diagnostics.
- Public Repositories: Large-scale data from Kaggle, Amazon Open Data, and government databases to provide variety and scale.
The modules specifically emphasize the vital but often overlooked aspects of the data lifecycle: outlier detection, handling missing values, and signal filtering.
Preliminary Results and Future Work
As a work in progress, the project has successfully developed modules for data preparation, visualization, clustering, and classification. These have been tested for cross-platform portability, ensuring they meet the software preferences of various engineering departments.
The next phase of the project involves a pilot rollout across six courses at Auburn University, spanning Chemical, Electrical, Computer Science, and Mathematics disciplines. The researchers plan to evaluate the impact of these modules using the Metacognition Awareness Inventory (MAI). This will determine if navigating complex data improves students' ability to monitor and regulate their own learning processes. Ultimately, the goal is to provide a "plug-and-play" repository of modules for the broader engineering education community through the ASEE network.
Summary updated