With projects spanning an array of industries, data types, and methodological approaches, the Master of Science in Analytics (MScA) program’s spring 2021 Capstone Showcase featured a field of twenty-four impressive teams, four of which were honored with Best in Show awards.
“The quality of all the work was truly impressive and exceptional,” said Greg Green, Executive Director for the MScA program. “The common element among the teams was a sense of professionalism, pride, and their commitment to excellence. I was especially thrilled to see that a number of alumni joined the presentations from around the globe, and I look forward to many of this year’s graduates staying connected.”
As the culmination of the student experience in the MScA program, the Capstone Project unites the academic lessons learned in the classroom with a real-world problem. The students lead all aspects of the project with the support of faculty, staff, and subject matter experts. Spring 2021’s projects and presentations gained praise in particular for their leading-edge analytic solutions and their effective storytelling and use of data visualizations.
The members of the Best-in-Show teams and their projects were
- Taylor Rasley, Taylor Williams, and Chris Lowe for “Multi-Family Water Fixture Classification Through Sound Data”
- Yun Huang, Duo Zhou, and Jingyun Jiang for “Failure Detection for Pharmaceutical Cold Chain Logistics”
- Laura Burns, Jason Rosenblum, and Josh Bender for “Forecasting Application for Labelmaster Sales” and
- Justin Cox, Drew Jacobs, and Roquiya Sayeq for “Simultaneous Localization and Mapping of Autonomous Robot”
Taylor Rasley, Taylor Williams, and Chris Lowe worked with Conservation Labs, a company that uses machine learning tools to enable cost-effective and sustainable water use. The team’s goal was to gain insight into water consumption as a way to help property managers conserve water and lower costs. Advised by Anil Chaturvedi, the team used a dual microphone sensor placed on water pipes to identify water usage through sound.
“Our team created a predictive classification model that used the sound data to identify the specific fixture associated with each water usage event,” they explained. “Through advanced sound event detection models, specifically convolutional neural networks and boosted tree models, we arrived at a solution that will help Conservation Labs deliver valuable new water usage insights in the multi-family residential property domain.”
As a cold-chain solution company that supports pharmaceutical transportation, Envirotainer maintains the temperature of cargo containers while delivering temperature-sensitive pharmaceuticals. Advised by Ashish Pujari, Yun Huang, Duo Zhou, and Jingyun Jiang developed a failure risk profile based on Envirotainer’s historical data to understand the characteristic failures that arise from temperature deviations, performance anomalies, and container parts malfunctions.
“We used different machine learning methods to analyze the data and predict the likelihood of failure prior to shipment,” they said. “With the recommendations and insights generated from our predictive analyses, Envirotainer can now reduce shipment failure by taking preventative measures and improving container handling training.”
Laura Burns, Jason Rosenblum, and Josh Bender, who were advised by Don Patchell, developed a prediction-based model for a core product line at Labelmaster, a company that provides software, products, and services for the shipping of dangerous goods. Their solution enhanced the company’s existing framework for estimating budget, production, and labor needs.
“The executive team at Labelmaster realized they had an opportunity to develop a data-driven solution to accurately forecast sales,” they explained. “We used Labelmaster’s historic monthly sales data, along with U.S economic indicators and freight market indicators, to predict more than twelve months of future sales for a key product line.”
Finally, Justin Cox, Drew Jacobs, and Roquiya Sayeq developed a method that gives autonomous robots the ability to navigate environments without GPS or pre-mapped paths. Advised by Yuri Balasanov, they used the Simultaneous Localization and Mapping (SLAM) technique to map and navigate an environment in real-time.
“Our Kalman filter implementation monitors data streaming from a robot’s onboard sensors,” they said. “From there, we used landmarks and the robot odometry to estimate the robot’s position within the environment as a way to satisfy localization and mapping.”
All the projects exhibited at the MScA’s Spring 2021 Capstone Showcase were roundly congratulated by the judges and the subject matter experts on hand. Praised for their creativity when it came to seeking out novel tools and technologies, students were also applauded for engaging their data sets on the deepest levels. Having risen to the challenge of steering their projects to completion, the MScA graduates have proof that they are now ready to tackle whatever analytics problems the world outside the classroom presents.