DHF Guide 7: Comprehensive Analysis of Rubrics and Automated Feedback

🔍 Guide 7: Comprehensive Analysis of Rubrics and Automated Feedback

This guide helps you break down exactly how the grading engine reads your submissions, how to interpret feedback phrases, and how rubrics vary between tracks. 

1. Understanding Lab Grading vs. Non-Lab AI GradingThe platform utilises two entirely different mechanisms to evaluate your work depending on the project type:

• Lab Project Grading Mechanics: For internal lab tasks (like Linux or VS Code tasks), grading is not done by an AI. Instead, a strict technical backend script directly checks your code for hard technical evidence. It checks things like whether a specific library was imported or if a precise directory loop script was written exactly as specified. 
  
  
• Non-Lab Project AI Grading Mechanics: For external uploads (like videos, documents, or repository links), a trained Artificial Intelligence engine reviews your submission against a structured framework. The AI reads text files, scans your repository files, or processes public links to find elements that match the instructor’s scoring sheet. 
  
  

2. Deconstructing the Criteria ChecklistWhen your grading results load, you will see specific score points broken down across multiple technical elements. For example, in a design track like UI/UX, the system breaks your work into distinct categories such as App Overview, User Flow, UI Evaluation, and UX Evaluation. In a development or data track, it might look for things like the Random Module, While Loops, or Conditional Logic. 

• Partial Passes: You do not need to get a perfect score to see feedback. The system can grant a partial pass if you fulfill basic requirements (such as creating the file in the right location) while flagging more complex programming requirements as failed if they are missing. 
  
  
• Decoding AI Feedback Phrases: If your submission does not line up with the assignment requirements, the AI grader will tell you exactly what is wrong in the text feedback window. For example: 
  
  
  • "Your submission didn't come through with the test or images... please upload your analysis..." This means you uploaded a completely unrelated or corrupt file that doesn't match the project parameters. 
    
    
  • "The repository contains a README.md and intro.txt file, neither of which include the required implementation..." This indicates that while your link worked and is public, you only uploaded text descriptions instead of actual programming code scripts. 
    
    
  • "The heavy ball animation is missing entirely, so a significant portion of the rubric will not be evaluated..." This tells you that you only completed half of the project prompt, and the grader had to skip grading the missing elements entirely. 
    
    

By carefully expanding the grading tab, reading these precise notes, and adjusting your files to match the missing items, you can systematically improve your score on subsequent attempts.