Innovative Uses in Education: William and Mary

Professors at William and Mary utilize the Virginia ODP to empower students to interpret, analyze, and communicate data findings.

The College of William and Mary logo

Empowering Students One Dataset at a Time

At ODGA, we are committed to empowering Virginia's citizens by putting data at their fingertips, enabling them to conduct meaningful analysis and gain deeper insights into the Commonwealth through a wide range of datasets and topics. That’s why we offer the Virginia Open Data Portal— a comprehensive platform featuring over 13,000 datasets from federal, state, and local sources for Virginians to explore and utilize.

 

Two professors at  William and Mary’s Raymond A Mason School of Business have utilized the Open data Portal as an opportunity for learning in their Machine Learning and Database Classes. In these classes, datasets on the portal are assigned for important projects geared towards data cleaning, data analysis, and interpretation and communication of data findings.

 W&M Database and Machine Learning Classes

In Professor Rodriguez Abitia’s database class, students are tasked with selecting a topic and sourcing relevant datasets from the Virginia Open Data Portal to support their research and analysis. He then asks students to consolidate data from multiple sources, clean it, and put the data all into the same format. Then students are tasked with looking for missing values and other data quality issues. After the data is cleansed and consolidated, the students are then instructed to build a previously designed data warehouse by connecting Alteryx with MySQL to automatically create the tables and feed them with the corresponding data. Students then focus on decision making with the data and create a dashboard using Tableau. They answer prompts with the data that they create using the dashboard and their analysis.

In Machine Learning 1, Professor Tremblay integrates the Open Data Portal as a resource for students to develop predictive models using real-world data. Students perform data cleaning, exploratory data analysis (EDA), feature engineering, and model training in DataCamp Workspaces, applying techniques such as linear regression, logistic regression, and k-nearest neighbors. Cross-validation and evaluation metrics like accuracy, precision, recall, and mean squared error (MSE) are used to assess model performance.

 

Beyond technical execution, students present their models and insights to our executive partners, receiving valuable industry feedback. This collaboration enhances both technical skills and business communication, bridging the gap between data science and real-world decision-making.

 

The project concludes with a poster presentation, summarizing key findings, challenges, and recommendations, preparing students for practical machine learning applications​.

 

Picture Caption: An example of one of the student extracts completed in the Database class.

Example of work with ETL processes and Dashboard creation a student completed in one of the classes mentioned
Example of a presentation given in the classes mentioned above.

Skills Gained: Technical and Communication

Students learned many technical skills through these classes, and took away a soft skill element as well. Both Professors Rodriguez Abitia and Tremblay mentioned that students needed to communicate their data findings in a way that non-technical audiences can understand. They had to use critical thinking to understand their audience and speak the audience’s “language.”

Students demonstrated both their data analysis and data communications skills throughout the entire project. One William and Mary student commented on their project related to food insecurity in Virginia: 

"Our project aimed to leverage machine learning to analyze and predict food insecurity trends in Virginia, providing a data-driven approach to identifying vulnerable regions. While our model is not designed for long-term forecasting, it offers valuable insights for regions with limited data, enabling better resource allocation. Addressing food insecurity requires a strategic approach, and we hope our work sparks meaningful discussions on how data can be used to guide impactful interventions both locally and globally."

ODGA encourages professors and educators across the Commonwealth to take full advantage of the Virginia Open Data Portal as a dynamic, growing resource for class projects, assignments, and lesson planning. By integrating this valuable tool into their curriculum, we can help broaden the minds of Virginia’s students and empower them with essential skills in data literacy and analysis.