Documentation

Component Breakdown

Statistical Methodology

Technology Stack

Database Schema

model Profile {
  id                    String   @id @default(cuid())
  userId                String?  @unique
  username              String   @unique
  avatarUrl             String?
  bio                   String?
  location              String?
  company               String?
  blog                  String?
  hireable              Boolean  @default(false)

  // Core Metrics
  score                 Float?
  percentile            Int?
  totalCommits          Int      @default(0)
  totalRepos            Int      @default(0)
  totalStars            Int      @default(0)
  totalForks            Int      @default(0)
  totalPRs              Int      @default(0)
  mergedPRs             Int      @default(0)
  openPRs               Int      @default(0)

  // Activity Metrics
  currentStreak         Int      @default(0)
  longestStreak         Int      @default(0)
  averageCommitsPerDay  Float    @default(0)
  mostActiveDay         String?
  weekendActivity       Float    @default(0)

  // Social Metrics
  followersCount        Int      @default(0)
  followingCount        Int      @default(0)
  organizationsCount    Int      @default(0)
  gistsCount            Int      @default(0)

  // Collaboration Metrics
  totalIssuesOpened     Int      @default(0)
  totalReviews          Int      @default(0)
  totalContributions    Int      @default(0)
  totalWatchers         Int      @default(0)
  totalOpenIssues       Int      @default(0)

  // Repository Health
  averageRepoSize       Float    @default(0)
  accountAge            Float    @default(0)
  accountCreatedAt      DateTime?

  // Language Data (JSON)
  languages             Json     @default("{}")
  frameworks            Json     @default("{}")

  // Repository Data (JSON array)
  topRepos              Json     @default("[]")

  // Contribution Data (JSON array)
  contributions         Json     @default("[]")

  // Scoring Components (JSON)
  scoreComponents       Json?
  scoringMethod         String?  // "fallback" or "pro"
  scoreStrengths        String[]
  scoreImprovements     String[]

  // Cache Management
  scannedAt             DateTime @default(now())
  lastLanguageScan      DateTime?
  lastFrameworkScan     DateTime?
  lastOrgScan           DateTime?

  // Indexes for performance
  @@index([username])
  @@index([score])
  @@index([scannedAt])
}

Data Flow Architecture

Security & Bot Protection

Analyzes a GitHub username and returns comprehensive developer metrics. Implements 24-hour caching and rate limiting.

{
  "username": "torvalds",
  "_honeypot": "",           // Must be empty (bot detection)
  "_timestamp": 1704067200000 // Page load time (timing validation)
}

{
  "success": true,
  "cached": false,
  "profile": {
    "username": "torvalds",
    "score": 96.93,
    "percentile": 97,
    "totalStars": 223690,
    "totalForks": 60864,
    // ... additional metrics
  },
  "nextScanAvailable": "2026-01-15T20:11:50.546Z",
  "hoursRemaining": 23
}

const response = await fetch('/api/analyze-username', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({
    username: 'torvalds',
    _honeypot: '',
    _timestamp: Date.now() - 2000 // 2 seconds ago
  })
});

const data = await response.json();

if (data.success) {
  console.log(`Score: ${data.profile.score}/100`);
  if (data.cached) {
    console.log(`Cached data, next scan in ${data.hoursRemaining}h`);
  }
}

Retrieves cached profile data for a given username. Fast endpoint (~50ms) for displaying dashboard data.

{
  "user": { "plan": "FREE" },
  "profile": {
    "username": "torvalds",
    "score": 96.93,
    "percentile": 97,
    "scoreComponents": {
      "impact": { "score": 99.99, "weight": 35, "source": "statistical" },
      "codeQuality": { "score": 38.37, "weight": 30, "source": "statistical" },
      "consistency": { "score": 72.96, "weight": 20, "source": "statistical" },
      "collaboration": { "score": 69.50, "weight": 15, "source": "statistical" }
    },
    "scoringMethod": "fallback",
    "totalStars": 223690,
    "totalRepos": 10,
    "languages": { "C": 98, "Rust": 0.3, "Shell": 0.4 },
    "topRepos": [ /* array of repository objects */ ],
    "contributions": [ /* array of contribution data */ ],
    // ... all profile fields
  }
}

const response = await fetch('/api/profile?username=torvalds');
const data = await response.json();

console.log(`Score: ${data.profile.score}/100`);
console.log(`Impact: ${data.profile.scoreComponents.impact.score}%`);
console.log(`Total Stars: ${data.profile.totalStars.toLocaleString()}`);

Calculates a user's global ranking position among all analyzed profiles. Real-time calculation using Prisma aggregation.

{
  "rank": 3,
  "totalProfiles": 1247,
  "percentile": 99.76,
  "score": 96.93
}

// Count profiles with higher scores
const higherScores = await prisma.profile.count({
  where: { score: { gt: userScore } }
});

// Rank is 1-based
const rank = higherScores + 1;

// Calculate percentile
const percentile = ((totalProfiles - rank + 1) / totalProfiles) * 100;

const response = await fetch('/api/global-rank?username=torvalds');
const { rank, totalProfiles, percentile } = await response.json();

console.log(`Rank #${rank} of ${totalProfiles}`);
console.log(`Top ${percentile.toFixed(2)}% globally`);

How accurate are the scores?

Scores are statistically accurate relative to our baseline population of 100,000+ developers. The system uses z-score normalization, which means a score of 70 always represents "better than 70% of developers" regardless of when it was calculated. However, scores reflect GitHub activity patterns, not developer skill, work ethic, or professional competence.

Why is my score lower/higher than expected?

The scoring system weighs impact (35%) most heavily. If you maintain popular open-source projects with many stars and forks, you'll score higher. Conversely, having many private repositories or working on closed-source projects won't increase your score since GitCheck only analyzes public data.

Common reasons for lower scores: few public repositories, low star count, infrequent commits, or a new GitHub account (account age affects several metrics).

How often can I re-analyze my profile?

Profiles are cached for 24 hours to reduce GitHub API usage and prevent abuse. After 24 hours, you can request a fresh analysis. The cache warning on your dashboard shows the next available scan time.

Does GitCheck access private repositories?

No. GitCheck only analyzes publicly available GitHub data. We never access private repositories, require OAuth authentication, or store sensitive information. All data comes from GitHub's public API endpoints.

How can I improve my score?

Focus on the four component areas: Impact (create valuable open-source projects that earn stars), Code Quality (maintain repositories consistently, close issues), Consistency (commit regularly, build streaks), and Collaboration (contribute PRs, do code reviews, join organizations).

Can I remove my profile from the database?

Yes. Contact us via GitHub issues on our repository or through the homepage contact information. We'll honor deletion requests within 7 days. Note that all data stored is already publicly available on GitHub.

Why does analysis take 30-45 seconds?

Full analysis requires fetching data from multiple GitHub API endpoints (repos, commits, PRs, contributions), calculating statistical metrics, and writing to the database. We use GraphQL to optimize this, but GitHub's API has inherent latency. Cached responses are served in ~50ms.

How is global ranking calculated?

Global ranking counts how many profiles in our database have a higher score than yours. If your score is 85.5 and 42 profiles score higher, your rank is #43. Percentile is calculated as: ((totalProfiles - rank + 1) / totalProfiles) × 100

Is this an official GitHub product?

No. GitCheck is an independent analytics platform that uses GitHub's public API. We are not affiliated with, endorsed by, or sponsored by GitHub, Inc.