About QUEST Data Analysis Projects

The applied data analysis course within QUEST teaches students how to develop insights from data that is generated by real-world companies. Accordingly, students are expected to address data quality issues that span from data ingestion, cleansing, and wrangling before they start analyzing the data. 

Success is achieved when the students can learn about data science in the real world and companies can derive value from the engagement. The faculty for the course ensure that students are learning the course material through lectures, case studies, and guest speakers. As the semester progresses, corporate partners often glean insights from students examining the data with a fresh set of eyes. Often, students will ask questions about the data collection and dissemination processes within companies that help companies improve their efforts to understand and leverage their data in a timely way. At the end of the semester, student teams will share their final deliverable with actionable recommendations for the client.

Data Analysis Project Examples

Spectrum Foods — Spring 2019

The Opportunity

Spectrum Foods is a distributor of poultry, beef and pork and other food products and non-food products for restaurants in the Mid-Atlantic Region. Since these restaurants require fresh supplies and have very little inventory on hand, Spectrum’s distribution to them is a critical part of their success. Although Spectrum has a great deal of data about their customers and their purchases, much of their supply chain model is built on giving their customers what they want with short turnaround time. The team had the opportunity to look at customer purchases to help predict future orders from their customers.

Data Science Methods

  • Data roll-ups across SKUs and time for each customer
  • Descriptive statistics and data trend analysis to understand sources of variability and drift in customer ordering
  • Clustering of data to help understand similar customer patterns across customers
  • Predictive modeling of customer orders within customer clusters

Recommendations

The team was able to help Spectrum Foods understand how some of their existing practices of providing good customer service and how some exceptions can better be anticipated in the future.

What The Client Said

“The QUEST data analysis team helped validate some processes that are already in place at Spectrum and opened my eyes to some areas of improvement. I would definitely be interested in working with another QUEST team in the future." — Josh Fanaroff, Chief Operating Officer, Spectrum Foods

Unilever — Fall 2018

The Opportunity

Unilever assembles final products based on a production run that brings together a build of materials from different parts of the supply chain. Since all the materials may not arrive of sufficient quality, Unilever has built in a safety factor in their build of material orders based upon the design of the system. Although the assembly and supply chain systems improve over time, the safety factor remains constant. Students were asked to analyze the number of parts left over from many production runs to help update the safety factor numbers.

Data Science Methods

  • Data joins across multiple Excel spreadsheets within Unilever’s systems
  • Data calibration and cleansing
  • Analysis of variance
  • Multivariate regressions to help isolate independent variables affecting variance in the number of leftover parts
  • Cost-benefit analysis of the recommendations

Recommendations

Students suggested updated safety factor numbers across different assembly systems and associated build of materials. The recommendations included estimates of the probability of excess or insufficient build of materials and the cost of increasing or decreasing the safety factors.