Analytics Students Present Project Results to Employers and Others
Over the summer, Master's of Business and Science Students in the Advanced Analytics Practicum worked on a wide variety of projects, including two that were sponsored by Janssen Pharmaceuticals, which is part of the Johnson & Johnson family of companies. For companies who may be interested in partnering with this class in the future, please reach out to Dr. Christie Nelson at email@example.com or Jennifer Cleary at firstname.lastname@example.org.
Above: MBS Student Sarah Yang explains her team's work on a predictive analytics model designed to help Janssen Pharmaceuticals better track the factors that affect the length of clinical trials.
According to Professor Christie Nelson, one of the co-Instructors for the course, "The projects covered a range of topics such as predictive analytics, natural language processing, regression analysis, etc. The students worked very hard over the summer to learn about their application areas, gathering and cleaning their datasets, first doing descriptive analytics to understand their data, and later modeling their data using both supervised and unsupervised approaches to come up with many interesting results. Not only did they work on their own projects, but they supported and offered ideas to their peers, letting them be exposed to many different types of research projects.”
Some of the projects featured during the poster session included:
Impact of weather condition and other factors on railroad accidents
By: Arindham Roy Chowdhary and David Weuste
From their data analysis, the team discovered that from 1976 to 2015, the most common type of railroad accidents were “Sun Kinks”. They occur when railroad tracks change shape, which alters the distance between rails, further causing engines to derail. Arindham points out the main cause of “Sun Kinks” is due to high temperature during summer season between 12pm to 6pm. The team also found these accidents happened all over the United States.
Crime Prediction in the City of Chicago
By: Dera Okafor and Winston James
The team predicted the class of a crime in the city of Chicago based on time, location, and related factors. Dera and Winston pointed out their data analysis discovered overall more day time crime vs night time. Theft was one of the top one on the list during the day time. They also pointed out certain community area were more crime prone than other areas.
Building Predictive Models for Phase I-IV Janssen Clinical Trial Lengths
By: David Itenberg, Yajie Zhang and Gaurav Sharma
This team, along with another, worked with leaders at Janssen Pharmaceuticals to better understand factors affecting the length of clinical trials. Janssen provided the team with cleaned data to do the analytics and Janssen managers were on hand to guide the students in the research. While the results of this analysis are confidential, David and Yajie mentioned that they were somewhat surprised by some of the results. Talking with Janssen employees, though, helped them to better understand the significance of their findings.
The two teams that partnered with Janssen also did a poster session for Janssen employees on-site at the company. Students were able to meet with the members of the clinical trials data team they assisted, as well as a wide range of other workers.
This experience challenged MBS students to reframe their poster presentations to focus more on the business challenges the models addressed, as well as the business relevant outcomes.
Above: (From left to right) Sam Mathew (Portfolio Analytics & Capacity Management Group Leader, Finance/PMO, Janssen Pharmaceuticals), Sarah Yang (MBS Student), Charlie Terng (MBS Student), Brent Myers (MBS Graduate), Chris Maier (Analyst, Statistical Modeling, Janssen), Christie Nelson (Professor of Analytics, MBS Program)
The MBS students told Jannssen executives that they learned a great deal from the interaction and they have a much clearer understanding of how to apply analytics methods to key business problems, as well as how best to discuss findings with executives and non-technical employees.