CMSE 495
Michigan State University Data Science Capstone.
Anticipating Challenges
Agenda (80 Minutes)
- (10 minutes) Example Case Study
- (40 minutes) Group Case Studies
- (20 minutes) Team Charter Review
- (10 minutes) Sharing with the class
Reminder: Team 3x3 Reports due every Monday (starting today)
Example Case Study: Not Enough Training Data
For this exercise your groups will review one of the case studies in your handout. However, to get some practice, you instructor will lead the entire class in the first case study to give you an idea what you should be doing.
Group Case Studies
As a group, spend about 30 minutes to review the handout of the remainder of the case studies. You should have time to get through about three (3) of them (more is fine). We don’t want all groups to do the same studies so start with the case study in the table below. Spend about 10 minutes per case study by doing the following:
- Pick someone to keep track of time
- Pick someone to take notes
- Pick someone to read the case study out loud and have a discussion
- Discuss the questions at the end of the case study.
- Focus on the question “What could be added to your team charter that may help avoid this type of problem?”
| Team | Case |
|---|---|
| HAP - Synthetic Data Generation | 2. Case Study: Visit from the CEO |
| City of Grand Rapids - Social Impact | 3. Case Study: Conference Presentation |
| ICER - User Data Analytics | 4. Case Study: Team Conflict |
| Intramotev - Automated Video Data Labeling for Autonomous Trains | 5. Case Study: Data Overload |
| Ford - Defect Prediction | 6. Case Study: What is the best accuracy? |
| Techsmith - Healthy and Engaged User Data Exploration | 7. Case Study: Not Enough Training Data |
| Tribal Start Program - Tribal Early Childhood Research Data | 8. Case Study: Working Alone |
| TwoSix - LLM to Graphs | 2. Case Study: Visit from the CEO |
| HFH - Revenue Cycle Prediction | 3. Case Study: Conference Presentation |
| CEPI - Anomaly Detection | 4. Case Study: Team Conflict |
| QSIDE - SToPA: MultiTown Data Analysis | 5. Case Study: Data Overload |
| Kellanova - Point of Sale Analysis | 6. Case Study: What is the best accuracy? |
Team Charter Review
This is a team paring exercise. You review the team charter for the other’s team (see table below). Have someone from the team that wrote the team charter present it to the group (10 minutes). Discuss key features and how you try to anticipate challenges for the semester. Have the group brainstorm ideas to make the team charter even better. Type up your feedback to the other group.
| Team A | Team B |
|---|---|
| Tribal Start Program - Tribal Early Childhood Research Data | HFH - Revenue Cycle Prediction |
| Techsmith - Healthy and Engaged User Data Exploration | HAP - Synthetic Data Generation |
| QSIDE - SToPA: MultiTown Data Analysis | Ford - Defect Prediction |
| ICER - User Data Analytics | City of Grand Rapids - Social Impact |
| Kellanova - Point of Sale Analysis | Intramotev - Automated Video Data Labeling for Autonomous Trains |
| TwoSix - LLM to Graphs | CEPI - Anomaly Detection |
Sharing with the class
In the last 10 minutes of class your instructor will have groups share what they learned with the class. The important outcome will be ideas you think are good to add to the team charter to help anticipate challenges that may occur throughout the semester.
Introduce Project Schedule assignment
The next deliverable is the project plan and schedule. Teams will use this time to read though the assignment and work on getting ready for the next steps.
Written by Dr. Dirk Colbry, Michigan State University
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.