Practicum project

During their data science and business analytics practicum, students gain rich hands-on experience in the field while working with a partner organization or company to support the development of big data and analytics solutions. Student teams utilize state-of-the-art big data tools and platforms along with data science and analytics techniques to develop effective proof-of-concept solutions to assist partners with their big data initiatives.

Softura logo

A team of students joined Softura, a leading digital IT consulting and custom application development service provider, to assist one of its long-term OEM customers with their Internet-of-Things (IoT) efforts. The customer manufacturers construction, mining, and other heavy equipment that is equipped with sensors to collect telemetry data.

The project's objective was to develop a solution platform that can foster a deeper understanding of the telemetry data, the equipment's metadata, caution data, and other contextual data to infer actionable patterns.  The team used analytics to create descriptive and diagnostic models that will predict when a machine will break down, thus, sending an alert to the customer ahead of time. The team also contributed to the development of an enhanced dashboard for visualizing patterns and trends.

As part of the project, students contributed to the following tasks:

  • Pattern matching/analysis of the telemetry data (~300k assets)
  • Descriptive analytics for the machines with caution/alerts
  • Developed models for predictive maintenance based on telemetry data, sensor data, caution data and other meta-data/contextual data (mostly equipment prognostics)
  • Prepared code sets for distribution along with documentation



A team of students joined the City of Detroit Office of Innovation & Emerging Technology (IET), a team of developers who partner with City departments to prototype, evaluate, and deliver digital tools that solve civic problems and information challenges, to assist in developing a platform that leverages open data from various City of Detroit systems (e.g. restaurant inspections, 311 calls, permitting, etc) along with Yelp reviews combined with predicative analysis methods to identify factors that lead to restaurants and food establishments facing health code violations.

The project's objective is to promote the use of open data to help address and prevent public health issues like food borne illnesses and provide citizenry with useful information about the health establishments.  In addition, the creation of a prediction model to identify which establishments are most likely to face health code challenges will help Detroit Health Department inspectors prioritize inspections and proactively guide establishments in resolving food safety issues before they can affect customers.

As part of the project, students contributed to the following:

  • Suggested and implemented data transformation methods and code sets (e.g. cleaning data, casting data types, geocoding, etc) to enhance underlying attributes and improve usability
  • Design and development of a descriptive analytics dashboard
  • Developed machine learning models for predicting future inspection outcomes and visualizing findings
  • Automated processes for sustainable implementation and prepared documentation
  • Published work (code and toolsets) as an open-source repository on City of Detroit's GitHub