Master Classes and Innovation in Teaching

We focus on developing hard skills–for example, strong resume bullet points, such as causal modeling, hypothesis testing and machine learning–and soft skills–for example, design thinking and problem solving in teams.

You learn to use new digital and cloud-based tools, such as biometric sensors, Microsoft’s Azure Machine Learning Studio, and app-based rapid prototyping toolkits.

Classes include:

  • MGT 349/505: Data Analytics (since 2013): From data triangulation and transformation (“garbage in, garbage out”), and causal modeling (“correlation does not imply causation”) toward a first algorithm. (Download summary)
  • MGT 317: Creating Smart Products:  (since 2015): Add a “Fitbit,” “remote control,” and “learning” to your product to make it “smart”; test prototypes and your assumptions about customer behavior and business models. (Download summary)
  • MGT 378: Social Media Analytics (since 2018): From descriptive analytics (running Google Analytics) and predictive analytics (conducting A/B testing) to prescriptive analytics (building product recommendation engines). (Download summary)

Example: Smart Coffeemaker

“Did I switch the coffeemaker off?”

Our graduate students Sara Al-Mahdi, Wendy Wei Ti Wen and Vincent Zhe You have busy lives. They love coffee but don’t want to think too much about operating their coffeemakers.

During this time of great technological advances, they felt that many kitchen appliances, such as coffeemakers, have been been left behind. You can check order status and track packages online, but there isn’t a way to find out whether you switched off your coffeemaker. So the students decided to turn their coffeemaker into a smart product.

The video shows a phase of an agile development process that utilizes a rapid-prototyping toolkit. You can see the team as it prepares a test of a new customer journey using an interactive prototype. The prototype bears little resemblance with a coffeemaker, because there is nothing wrong with the brewing process itself. Instead the focus is on designing a new user experience and on testing user interaction alternatives using a real sensor (heat), actuators (LED and buzzer) and interaction logic (app on smartphone screen).

Example: Building Recommendation Engines

“How to sell more – and quickly?”

Recommendations from family, friends and customers remain the most credible form of advertising among consumers according to the market research experts at Nielsen. So why would any company let someone else solicit and curate customer ratings to provide product recommendations?

Yet, product ratings and user recommendations are typically left to third parties, such as Amazon or Yelp. Furthermore, nobody knows the product or service better than its maker. In fact, great effort and money is expended explicitly to differentiate an offering. No company should sell its products without recommendations.

Our graduate students Afnan Bawakid, Josh Griffith, Alexey Migerkin, and Thomas Slaughter decided to do better. As future product managers they decided to use machine learning to construct a recommendation engine as a group project.

The video illustrates how the team worked strategically as well as hands-on. In terms of strategy the team chose to manage the project using a simplified, 3-step data analytics process that corresponds well with CRISP-DM, a widely used open standard process for data mining applications. Hands-on development was done using Microsoft’s Azure Machine Learning Studio, a cloud-based development environment, which allowed the team to quickly run exploratory diagnostics and create a first recommendation engine prototype.