Judging Setup

​

Overview

​

Judging Data Model

​

Core Models

model Track {
  id             String           @id
  name           String
  dhYear         String
  ProjectTrack   ProjectTrack[]
  Table          Table[]
  RubricQuestion RubricQuestion[]
  
  @@unique([name, dhYear])
}

model Project {
  id             String          @id
  name           String
  description    String
  link           String          // Devpost URL
  tracks         ProjectTrack[]
  judgingResults JudgingResult[]
  TimeSlot       TimeSlot[]
  dhYear         String
}

model Table {
  id            String          @id
  number        Int
  trackId       String
  track         Track           @relation(...)
  dhYear        String
  TimeSlot      TimeSlot[]
  JudgingResult JudgingResult[]
  
  @@unique([number, dhYear])
}

​

Track Configuration

​

Common Track Examples

• Best Overall
• Best Use of AI
• Best Hardware Hack
• Best Healthcare Solution
• Sponsor Prize - Company X

​

Creating Tracks

await prisma.track.create({
  data: {
    name: "Best Use of AI",
    dhYear: "DH12"
  }
})

​

Project Import

​

CSV Upload Process

1. Navigate to /admin/judging
2. Upload CSV file in the Import Project Data section
3. System processes and creates Project records

​

Supported Platforms

Project Name,Description,URL,Track 1,Track 2
Awesome Project,Cool description,https://devpost.com/...,Best AI,Best Overall

​

CSV Processors

import { DevpostCSVProcessor } from "../utils/csvProcessors"

<CSVUploader csvProcessor={new DevpostCSVProcessor()} />

• Project records with name, description, link
• ProjectTrack associations (many-to-many)

​

Table Assignment

​

Creating Tables

await createTables({
  projectsPerTable: 10 // 10 projects rotate through each table
})

1. Counts total projects
2. Divides by projectsPerTable
3. Creates that many Table records
4. Assigns projects round-robin to tables

• 100 projects
• 10 projects per table
• Creates 10 tables (numbered 1-10)

​

Table Distribution

​

Timeslot Scheduling

​

Creating Schedule

await createTimeSlots({
  startTime: "2024-01-15T10:00:00Z"
})

1. Takes the startTime
2. Creates rotating schedule for all projects
3. Each project gets multiple timeslots (one per table rotation)
4. Allocates time based on number of projects and tables

​

Timeslot Model

model TimeSlot {
  id        String   @id
  startTime DateTime
  endTime   DateTime
  project   Project  @relation(...)
  projectId String
  table     Table    @relation(...)
  tableId   String
  dhYear    String
  
  @@unique([projectId, startTime])
  @@unique([tableId, startTime])
}

• Each project only scheduled once per timeslot
• Each table only has one project per timeslot

​

Schedule Output

• Total Duration: “3 hours and 45 minutes”
• End Time: “13:45”
• Number of Tables: “10”

• Plan judging duration
• Communicate schedule to participants
• Assign judge availability

​

Rubric Configuration

​

Rubric Question Model

model RubricQuestion {
  id        String           @id
  title     String           // Short title
  question  String           // Full question (supports Markdown)
  points    Int              // Max points (e.g., 10)
  trackId   String
  track     Track            @relation(...)
  responses RubricResponse[]
}

​

Creating Questions

1. Select a Track
2. Enter Title: “Technical Complexity”
3. Enter Question: “How technically challenging was the implementation?”
4. Set Points: 10
5. Click Create Question

​

Question Formatting

## Evaluate the following:

- Code quality and architecture
- Use of modern frameworks  
- Technical difficulty
- Innovation in approach

**Consider**: Is this pushing boundaries or using standard approaches?

​

Bulk Import

[
  {
    "title": "Technical Complexity",
    "question": "How technically challenging was the implementation?",
    "points": 10,
    "trackId": "clxxx"
  },
  {
    "title": "Design Quality",  
    "question": "How polished is the user interface and experience?",
    "points": 10,
    "trackId": "clxxx"
  }
]

​

Typical Rubric Structure

1. Technical Complexity (15 points)
2. Design & UX (10 points)
3. Impact & Usefulness (15 points)
4. Presentation Quality (10 points)

​

Judge Assignment

​

Assigning Judge Role

• Via admin role management interface
• Via QR scan at “judges” station

​

Judge Interface

1. View their assigned schedule (optional)
2. Select a project to judge
3. See the rubric questions for the track
4. Score each question
5. Submit their judging result

​

Judging Results

model JudgingResult {
  id String @id
  
  judgeId String
  judge   User   @relation("judge", ...)
  
  project   Project          @relation(...)
  projectId String
  responses RubricResponse[] // Individual question scores
  dhYear    String
  
  table   Table  @relation(...)
  tableId String
  
  @@unique([judgeId, projectId])
}

model RubricResponse {
  id              String         @id
  score           Int            // Points awarded (0 to question.points)
  judgingResultId String
  questionId      String
  judgingResult   JudgingResult  @relation(...)
  question        RubricQuestion @relation(...)
  
  @@unique([judgingResultId, questionId])
}

​

Leaderboard

​

Score Calculation

1. Sum all RubricResponse scores for a project
2. Average across all judges who scored it
3. Optionally filter by track

​

Leaderboard Display

​

Track Filtering

• All Tracks: Combined leaderboard
• Specific Track: Winners for that track/prize

​

Real-Time Updates

useQuery(getLeaderboard, {
  refetchInterval: 30 * 1000,
  refetchIntervalInBackground: true
})

​

Complete Judging Workflow

​

Pre-Event Setup

1. Create tracks for your prizes/categories
2. Import projects from Devpost CSV
3. Assign projects to tracks (done during import or manually)
4. Create tables based on venue layout
5. Generate timeslot schedule
6. Configure rubric questions for each track
7. Assign judge roles to external judges

​

During Judging

1. Judges arrive and scan at “judges” station (grants JUDGE role)
2. Judges access judging portal on their device
3. Judges visit tables according to schedule
4. Judges score each project using the rubric
5. Scores appear on leaderboard in real-time

​

Post-Judging

1. Review leaderboard for each track
2. Export results for prize announcements
3. Verify minimum number of judges per project
4. Handle edge cases (ties, low judge counts)
5. Announce winners at closing ceremony

​

Best Practices

​

Track Design

• 3-5 tracks max for small events (under 100 projects)
• Align with sponsors - create tracks for sponsor prizes
• Broad categories - avoid overly specific tracks

​

Table Layout

• 10-15 projects per table is ideal
• Physical space - ensure tables fit in venue
• Power outlets - each table needs power

​

Judging Schedule

• 5-7 minutes per project for judge scoring
• Buffer time between rotations for movement
• 3+ judges per project for reliable averages

​

Rubric Design

• 4-6 questions per rubric
• Clear criteria - avoid subjective questions
• 50-100 total points for granularity
• Track-specific - different rubrics for different tracks

​

Judge Management

• Brief judges before event on rubric usage
• Distribute evenly - ensure each project gets similar judge count
• Monitor progress - check leaderboard to identify under-judged projects

​

Troubleshooting

​

Projects Missing from Schedule

• Verify project was imported successfully
• Check that dhYear matches current year
• Ensure project is assigned to at least one track

​

Duplicate Timeslots

• Unique constraints prevent this, but if it occurs:
• Delete all timeslots and regenerate
• Verify table records are correct

​

Judge Can’t Access Portal

• Verify user has JUDGE role
• Check that dhYear is set correctly
• Ensure judge is logged in

​

Leaderboard Shows Incorrect Scores

• Scores are averaged across judges
• Check that all responses have correct questionId
• Verify points awarded are within question.points limit

​

Related Pages

• User Management - Assign JUDGE role
• Admin Dashboard - Configure dhYear
• Event Management - Judge check-in process