Month: December 2022

Atlas Air Worldwide

Atlas Air Worldwide is a leading global provider of outsourced aircraft and aviation operating services. From shipping precious perishables or heavy construction equipment to arranging large-group passenger charters for celebrities or dignitaries, Atlas Air’s technologically advanced fleet of 747s, 777s, 767s and 737s offers flexible, creative and award-winning solutions that meet our customer’s unique needs – quickly, safely and reliably.


Atlas partnered with UConn as an outgrowth of the Company’s Elevate Initiative, an organization-wide movement to identify, innovate and facilitate opportunities to transform Atlas’ employee experience. As part of their capstone project, 37 graduate students in the Business Analytics and Project Management program were asked to develop and propose solutions to help optimize flight schedules – a significant factor impacting the quality of work life for crewmembers at Atlas. The students were challenged to find insights that would meet the requirements of Atlas customers, who look to the airline for flexibility, while taking into account crewmembers’ needs for consistency and predictability. After analyzing and compiling fully de-identified flight data records, the students presented their findings and final recommendations.

“These were some bright young people who were interested in Atlas and in the air freight business as a whole, and it really showed in their work,” said Scott Lindsay, Vice President, Crew Planning and Analysis. “They had some really good suggestions and ideas about how to approach the complex data available to us. The Company is reviewing those ideas with a plan to incorporate some of them – or elements of them – into future improvements. This was a very important exercise in our on-going efforts to improve scheduling, and these insights will help shape our thinking.”


It was a pleasure working with UConn, CABA and Prof Eigo for the capstone project. The students in the business analytics course were very creative, enthusiastic and resourceful as they went about using our product and the provided datasets to extract insights from marketing-related datasets for our customers.

Our mission at Unscrambl is to simplify access to data and insights for both technical and non-technical business users using natural language-based interfaces. At the time we started working on the capstone, we had released a couple of new features, especially around automated data story-telling, along with an emphasis on providing insights on marketing use-cases. The students in the course were given the challenge of using our conversational AI product for exploring marketing datasets from multiple customers, creating stories from these datasets, and coming up with story templates that we could incorporate in our product.

We gained a lot of value from the project, both in terms of ideas for marketing data story templates as well as in terms of valuable insights on the usability of our product. The students also provide great feedback on how our product could be further improved in terms of new user on-boarding, analytics and collaboration features. The students were also given the opportunity to present their findings to our customers, which they did a great job in. We hope that they gained valuable experience in uncovering and presenting stories on real-world datasets.

Anand Ranganathan
Chief AI Officer


Clorox LogoClorox disinfecting wipes are the star products of Clorox, having a lion’s share of sales amongst its competitors. The company has been traditionally using just the past sales data for forecasting its immediate future demand. The major problem with these kinds of models are they do not take into account some of the very influential factors which strongly effects the sales, creating unknown pulse periods. This causes the inventory to either come short of the market demand , resulting in losing the existing customers to competitors or exceeding the market demand, which may result in over stocking the inventory. Both of which leads to loss.

By using sophisticated models which could correctly predict the sales due to influential factors, Clorox could clearly handle their inventory better and reach the market in a timely manner. Identifying the pulse periods will give the company a competitive edge in creating new marketing strategies to increase market demand. In addition, rolling out inventory at the opportune time assists in customer retention and increases overall sales.

  • Analyze the factors affecting the sales of disinfecting wipes
  • Determine the drivers behind pulse periods
  • Build model based on previous sales data
  • Predict the unit sales for the year 2016


Priceline partnered with UConn to sponsor a capstone project for second year students in the MSBAPM program. 

Priceline Logo

Students directly applied their academic knowledge to identify how Priceline can better target the more than 2 billion promotional emails sent every year. After analyzing a sample of anonymized search and booking data, students presented final recommendations to Priceline’s product analytics and marketing teams.

Basement Systems

Seeking insights into the home improvement market, Basement Systems Inc. ( partnered with the UConn Business Analytics and Project Management program on several inter-related projects over the course of two years. A division of Contractor Nation, Basement Systems is headquartered in Seymour, Connecticut and comprises a nationwide network of home improvement Basement Systems Logocontractors offering basement waterproofing, foundation repair, insulation, mold prevention, and radon mitigation.

Graduate students in the program analyzed externalities that might affect demand for home improvement services, looking at the effect of weather, energy costs, the real estate market, and other geographic, economic, and demographic data. Other analyses looked at the cost effectiveness of different marketing channels and other factors affecting a company’s growth, including social media presence, competition, and reviews.

In addition to data analyses, the students used text mining tools to explore customer sentiment as expressed in online reviews. “These insights into our customers are an important factor in understanding market dynamics,” explained Jenny Hui Liu, a graduate of the program and Data Analyst at Basement Systems who worked with the students on the project’s parameters.

Richard Fencil, head of the Treehouse Internet Group, ( the company’s in-house marketing agency, said the results of the projects were being used in a number of ways. “The insights were immediately helpful in mapping the customer journey, expanding weather-related advertising, and our ability to provide support to our contractor network through our year-round training programs. The more we understand the interplay between reviews, referrals and online research, the clearer it is that a contractor has to deliver exceptional service to grow its business.”

Pentation Analytics

Pentation Analytics is a Big Data Analytics company that enables insurers and intermediaries to better engage with policy holders. We provide predictive intelligence and process automation tools that address the core insurance use-cases of increasing retention, crPentation Analyticsoss-sales and optimizing claims. We are based out of Hartford, Connecticut and we are highly engaged with UConn’s Center for the Advancement of Business Analytics (CABA).

Pentation Analytics and CABA organized a hackathon focused on customer retention. Customer retention is a global business problem area in which much work is being done, but it is still not enough. The problem statement read “Insurance customer retention insights through dashboards”. Students were given P&C Insurance carrier data, both structured and unstructured. Student turnout was overwhelming and it was very interesting how the teams approached the problem. In addition, seeing data from their eyes was an amazing and enriching experience. In terms of the student experience, we hope inputs provided by our team during presentations will help these talented students to bridge the academic-Industry gap.

The hackathon was very successful and we look forward to many more fruitful endeavors between Pentation and CABA.

-Prahant Roy, Pentation Analytics

Aureus Analytics

Aureus Analytics was born to fulfill the need to do more with your data but with lesser tools. What started as a need to have access to real time insights at the point of decision, has now blossomed into a smorgasbord of highly energetic and restless individuals creating a great product to revolutionize customer experience in Insurance Industry.

After meeting with CABA and learning about the program structure and how CABA engages with industry for the benefit of the students as well as the industry, we were impressed. The endeavor to keep the engagements symbiotic was very visible and absolutely novel. In our CABA Data Challenge, our approach was to engage the students by introducing them to something new and exciting. We built the challenge around a customer experience paradigm and demonstrated our concepts and technology to them in an extensive two phase process. In phase I, the students were exposed to our application and its use via a simulated demo, and a simple quiz was then used to assess and select the suitable candidates for phase II. Phase II was more extensive. Direct access to our technologies and sample data was provided to them to evaluate the customer events and experience paradigm.

The commitment and feedback from the students was phenomenal and encouraging. They understood and critiqued the approach in a very constructive way. They also took the initiative and analyzed the data from a customer perspective. We are looking forward to selecting a few students for an internship to carry forward the engagement that was established with this Data Challenge. We sincerely thank the CABA team of Anna Radziwillowicz, Jennifer Eigo & John Wilson for all their help and support. We look forward to more similar engagements, projects and Data Challenges.
-Nitin Purohit, Co-founder & CTO

Capital Community College


Capital Community College Logo

Capital Community College worked with a team of UConn graduate students to help boost its fall-to-fall retention rate, which currently hovers around 45 percent.  The UConn students’ findings were based on five years of data supplied by CCC, and all student identification remained confidential. The results of this predictive analytics project outlined the characteristics of CCC students who are most likely to succeed, and those who are at greater risk of dropping out.

“We hear a lot of stories from individual students, but until now we did not have the aggregate findings from the data analytics. The value of this kind of research is immeasurable. We are grateful to UConn, our colleagues down the street…we are all stakeholders in the success of our community.” G. Duncan Harris, Interim CEO at Capital Community College

The UConn students, in their consulting role, offered a number of suggestions to CCC administrators, such as augmenting the offering of financial advising to students who aren’t eligible for aid, considering more flexible class times for those who are working, studying which advisers work best with specific student populations, and adding additional support for students in the higher-risk categories.

Administrators from both colleges are considering extending the partnership. UConn students would like to gather more data, such as how many hours the average student is working, whether marital status impacts retention, how dependents influence completion of college, and the most common reasons students leave CCC.

“This partnership is phenomenal,” said Miah LaPierre-Dreger, CCC’s Interim Dean of Academic Affairs. “A lot of community colleges across the country would love to have this kind of data about their programs.”

“You’ve affirmed that we’re on the right track with the recent student-success strategies we’ve deployed and provided a data-informed platform for the launching of others,” Harris said. “The spirit of collegiality and support for our shared community exhibited in this project was phenomenal.”

Voya Financial


Voya Financial Logo

Voya’s Investment Management business sends a wide variety of emails to current and potential distribution partners (consultants, advisors, fund analysts, etc.). These emails contain market commentary, product information, webinar invitations and many other topics intended to build engagement and add value.

Students were challenged to identify and quantify attributes that lead to engaging email communications from Voya.

Some potential insights:

  • What characteristics drive the success of individual emails?
  • What sequences of emails affect engagement?
  • When should emails be sent to drive the greatest engagement?

Student response to this first-time UConn-Voya partnership event was outstanding.

“It was a lot of fun working with the students and seeing what they came up with. They had a fresh perspective on real challenges we’re working on today. It was very satisfying to be able to talk through their thought process and engage with such a talented and motivated group.”
-Duncan Sanford, Analytics Director, Analytics Center of Excellence

“This analytics challenge was a great example of how we’re building partnerships with leading universities to create a pipeline of ideas and talent. Through internships and analytics projects like this one, we’re creating a buzz about Voya on campuses that’s helping to generate new insights as well as to compete for talent.”
-Tom Hamilton, VP, Analytics Center of Excellence


LIMRA asked the students to build a model to predict variable annuity cancellations, based on a large data set with dozens of variables. In parallel with our internal efforts, the students adeptly explored the data, built models, developed theories, and presented their recommendations. We used their results to confirm our findings, and their fresh perspective even gave us some new ideas. We arLimra Logoe planning to publish our findings this year, in no small part due to the students’ contributions.


Buyer / Non-Buyer Project
LIMRA asked the students to build a model to predict who was likely to buy individual life insurance, based on a consumer survey. Despite having a limited amount of data, the students were able to identify several personal characteristics and life events that make someone more likely to buy. Their work has helped us develop profiles of types of customers who are most likely to buy, which we are planning to publish this year.