IS480 Team wiki: 2016T1 MARKSmen Final
Jump to navigation
Jump to search
Project Progress Summary
Project Highlights
- Launched Team Question activity in IS303
- Conducted UAT with 3 SMU SIS professors
- Launched Mentoring activity in IS303
- Used Collaborative Submission activity for graded submissions in IS303
- Conducted UT IV with 25 SMU students in a simulated classroom setting
Project Challenges
- Making sure the project is perpetually ready for usage and deployment week in week out
- Coping with a large volume of user feedback
- Adapting product and sprint backlogs to fluctuating importance of requirements
Project Achievements
- Usage in 4 sections of SMU seminars after acceptance
- 105 Collaborative submissions
- 289 Team MCQ submissions
- 33 Team Questions asked
- 22 Mentor-Mentee pairings created
- 13 individual questions and responses in Questions (forum)
- 437 teams formed
Project Management
Project Status
Planned Project Schedule (Mid Terms)
Actual Project Schedule (Current)
Change Management
Project Metrics
Bug Metric
Feature Metric
Workload Metric
Technical Complexity
Mentoring Algorithm
Goal: To most efficiently and effectively allocate a mentee to a mentor. A mentor is a student who just completed an assignment, and a mentee must be a student who is considered "weaker" based off historical data.
Considerations:
- Allocation is based off previous assignment completion, and hence is a stochastic metric
- Allocation should not be static: Algorithm should be able to be re-run and not always return the same user (We want to eliminate inherent bias against a single student)
- Allocation must still take into consideration student strength and weaknesses. Allocation chances should be relative, not absolute
Solution: Fitness proportionate Selection
Step 1:
- Filter out participants who have completed the challenge or have already been allocated a mentor. The remaining array of participants is our sample.
Step 2:
- Identify the number of challnges each student has historically completed
Step 3:
- Assign a weight to each student based on the number of historically completed challenges.
- This attribute, called 'fitness', has an inverse relationship with the number previously of previously completed challenges.
- A higher fitness score indicates a higher percentage chance of being selected by the algorithm.
- Code and attribute transformation shown below.
Step 4: Run randomised selection on a normalised weighted object
- The array of participants must be transformed into a 'normalised object', with each key representing a range of values.
- A randomiser falling in this range of values would equate to the participants at that key being selected as the mentee.
- findRangeKey() performs un upward inclined selection of a key based on a random roll.
- In doing so, we have now selected a mentee based on their proportionate 'Weakness', which meets our the consideration critetion.
Team Formation Race Conditions
Recursive Promises
Quality of Product
Project Deliverables
Topic of Interest | Link |
---|---|
Project Management | Minutes |
Metrics | |
Risk Management | |
Change Management | |
Project Overview | Project Overview |
Team's Motivation | |
Project Documentation | Scenario |
Prototypes | |
Diagrams | |
Technologies | |
Testing | Testing Documentation |
Quality
Dynamic Height Expansion
Testing
Reflection
Team Reflection
Sponsor Comment(s)