goodgo-platform/docs/explorations/ANALYTICS_ARCHITECTURE.md

GoodGo Analytics Module - Architecture Exploration Summary

1. MODULE STRUCTURE (DDD Layers)

File Organization

apps/api/src/modules/analytics/
├── analytics.module.ts              # NestJS module (CQRS setup, DI)
├── index.ts                         # Exports
├── application/                     # Application layer
│   ├── commands/                    # Write operations
│   │   ├── generate-report/
│   │   ├── track-event/
│   │   └── update-market-index/
│   ├── queries/                     # Read operations
│   │   ├── batch-valuation/
│   │   ├── get-district-stats/
│   │   ├── get-heatmap/
│   │   ├── get-listing-ai-advice/
│   │   ├── get-market-report/
│   │   ├── get-nearby-pois/
│   │   ├── get-neighborhood-score/
│   │   ├── get-price-trend/
│   │   ├── get-project-ai-advice/
│   │   ├── get-valuation/
│   │   ├── industrial-valuation/
│   │   ├── predict-valuation/
│   │   ├── valuation-comparison/
│   │   ├── valuation-explanation/
│   │   ├── valuation-history/
│   │   ├── _shared/                 # Shared utilities
│   │   │   └── ai-json-client.ts
│   ├── event-handlers/
│   │   └── listing-created-moderation.handler.ts
│   └── __tests__/
├── domain/                          # Domain layer
│   ├── repositories/
│   │   ├── market-index.repository.ts
│   │   ├── valuation.repository.ts
│   ├── services/
│   │   ├── avm-service.ts
│   │   └── neighborhood-score.service.ts
│   ├── entities/
│   └── aggregates/
├── infrastructure/                  # Infrastructure layer
│   ├── repositories/
│   │   ├── prisma-market-index.repository.ts
│   │   └── prisma-valuation.repository.ts
│   ├── services/
│   │   ├── ai-service.client.ts
│   │   ├── http-avm.service.ts
│   │   ├── market-index-cron.service.ts
│   │   ├── neighborhood-score.service.ts
│   │   └── prisma-avm.service.ts
│   └── __tests__/
└── presentation/                    # Presentation layer
    ├── controllers/
    │   ├── analytics.controller.ts
    │   ├── avm.controller.ts
    │   └── index.ts
    ├── dto/                         # Request/Response DTOs
    │   ├── avm-compare-query.dto.ts
    │   ├── avm-explain-query.dto.ts
    │   ├── batch-valuation.dto.ts
    │   ├── get-district-stats.dto.ts
    │   ├── get-heatmap.dto.ts
    │   ├── get-market-report.dto.ts
    │   ├── get-nearby-pois.dto.ts
    │   ├── get-price-trend.dto.ts
    │   ├── get-valuation.dto.ts
    │   ├── industrial-valuation.dto.ts
    │   ├── predict-valuation.dto.ts
    │   ├── valuation-comparison.dto.ts
    │   ├── valuation-history.dto.ts
    │   └── index.ts
    ├── __tests__/
    └── index.ts

Architecture Pattern: DDD + CQRS

• Commands: State mutations (GenerateReportHandler, TrackEventHandler, UpdateMarketIndexHandler)
• Queries: Read operations with cache-aside pattern
• Event Handlers: Listen to domain events (listing-created-moderation)
• Repositories: Abstract data access (Market Index, Valuation)
• Services: Business logic (AVM, Neighborhood Scoring)

---

2. CONTROLLER & ENDPOINT STRUCTURE

Analytics Controller (analytics.controller.ts)

Routes: GET/POST /analytics/* and public endpoints

Key Endpoints:

GET  /analytics/market-report          → GetMarketReportQuery
GET  /analytics/price-trend            → GetPriceTrendQuery
GET  /analytics/heatmap                → GetHeatmapQuery
GET  /analytics/district-stats         → GetDistrictStatsQuery
GET  /analytics/valuation              → GetValuationQuery (by propertyId or coords)
POST /analytics/valuation              → PredictValuationQuery (manual input form)
POST /analytics/valuation/batch        → BatchValuationQuery (1-50 properties)
GET  /analytics/valuation/history/:id  → ValuationHistoryQuery
POST /analytics/valuation/compare      → ValuationComparisonQuery (2-5 properties)
GET  /analytics/neighborhoods/:district/score → GetNeighborhoodScoreQuery (no auth)
GET  /analytics/pois/nearby            → GetNearbyPOIsQuery (no auth, public)
POST /analytics/listings/:id/ai-advice → GetListingAiAdviceQuery (Claude)
POST /analytics/projects/:id/ai-advice → GetProjectAiAdviceQuery (Claude)

AVM Controller (avm.controller.ts)

Routes: GET/POST /avm/*

Key Endpoints:

POST /avm/batch                → BatchValuationQuery
GET  /avm/history/:propertyId  → ValuationHistoryQuery
GET  /avm/compare              → ValuationComparisonQuery (query string: ids)
GET  /avm/explain              → ValuationExplanationQuery (valuationId)
POST /avm/industrial           → IndustrialValuationQuery

Guard & Decorator Stack

@ApiBearerAuth('JWT')              // Swagger doc
@UseGuards(
  EndpointRateLimitGuard,          // Redis sliding-window rate limit
  JwtAuthGuard,                    // Verify JWT token
  QuotaGuard                       // Check user subscription quota
)
@RequireQuota('analytics_queries') // Decorator: specify quota resource
@EndpointRateLimit({
  limit: 10,
  windowSeconds: 60,
  keyStrategy: 'user' | 'ip'      // Rate limit by authenticated user or IP
})

---

3. QUERY/HANDLER PATTERN (CQRS Implementation)

Query Definition (Example: GetPriceTrendQuery)

export class GetPriceTrendQuery {
  constructor(
    public readonly district: string,
    public readonly city: string,
    public readonly propertyType: PropertyType,  // From @prisma/client
    public readonly periods: string[],
  ) {}
}

Handler Pattern (Example: GetPriceTrendHandler)

@QueryHandler(GetPriceTrendQuery)
export class GetPriceTrendHandler implements IQueryHandler<GetPriceTrendQuery> {
  constructor(
    @Inject(MARKET_INDEX_REPOSITORY) 
    private readonly marketIndexRepo: IMarketIndexRepository,
    private readonly cache: CacheService,         // Cache-aside pattern
    private readonly logger: LoggerService,
  ) {}

  async execute(query: GetPriceTrendQuery): Promise<PriceTrendDto> {
    try {
      const cacheKey = CacheService.buildKey(
        CachePrefix.MARKET_TREND,
        query.district,
        query.city,
        query.propertyType,
        query.periods?.join(','),
      );

      return this.cache.getOrSet(
        cacheKey,
        async () => {
          const trend = await this.marketIndexRepo.getPriceTrend(...);
          return { district, city, propertyType, trend };
        },
        CacheTTL.MARKET_DATA,  // 30 minutes
        'price_trend',         // Metric label
      );
    } catch (error) {
      if (error instanceof DomainException) throw error;
      this.logger.error(...);
      throw new InternalServerErrorException('...');
    }
  }
}

Key Pattern:

1. Inject repository (abstraction) and cache service
2. Build cache key with CacheService.buildKey()
3. Use cache.getOrSet() for cache-aside pattern
4. Specify TTL from CacheTTL constant and metric name
5. Catch DomainException separately, rethrow or wrap

---

4. REDIS CACHING PATTERNS

Cache Configuration (CacheService)

Cache TTLs:

CacheTTL = {
  LISTING_DETAIL: 300,           // 5 min
  SEARCH_RESULTS: 120,           // 2 min
  DISTRICT_STATS: 300,           // 5 min
  MARKET_REPORT: 900,            // 15 min
  HEATMAP: 300,                  // 5 min
  MARKET_DATA: 1800,             // 30 min (price trends)
  USER_PROFILE: 600,             // 10 min
  USER_QUOTA: 60,                // 1 min (frequently invalidated)
  PLAN_LIST: 3600,               // 1 hour
  REFERENCE_DATA: 86400,         // 24 hours (static districts/wards)
}

Cache Key Prefixes:

enum CachePrefix {
  LISTING = 'cache:listing',
  SEARCH = 'cache:search',
  GEO_SEARCH = 'cache:geo_search',
  MARKET_REPORT = 'cache:market:report',
  MARKET_TREND = 'cache:market:trend',        // For price trends
  MARKET_HEATMAP = 'cache:market:heatmap',
  MARKET_DISTRICT = 'cache:market:district',
  USER_PROFILE = 'cache:user:profile',
  USER_QUOTA = 'cache:user:quota',
  VALUATION = 'cache:valuation',
  PLAN_LIST = 'cache:plan:list',
  REFERENCE = 'cache:reference',
  AGENT_LISTINGS = 'cache:agent:listings',
}

Cache-Aside Implementation:

async getOrSet<T>(
  key: string,
  loader: () => Promise<T>,
  ttlSeconds: number,
  resource: string,  // Metric label
): Promise<T> {
  // 1. Fast-path: if Redis unavailable, call loader directly (graceful degradation)
  if (!this.redis.isAvailable()) {
    // Metric: cache_degradation_total[resource]++
    return await loader();
  }

  // 2. Try to get from cache
  // 3. If miss: call loader, store in Redis, return
  // 4. Metrics: cache_hit_total or cache_miss_total[resource]++
}

Metrics Tracked:

cache_hit_total[resource]          // Counter: hits by resource
cache_miss_total[resource]         // Counter: misses by resource
cache_degradation_total[resource]  // Counter: fallbacks when Redis down

Rate Limiting with Redis (Sliding Window)

Lua Script in EndpointRateLimitGuard:

-- Sliding-window algorithm using Redis sorted set
-- KEYS[1] = rate limit key (e.g., "rate:user:123:POST:/analytics/valuation")
-- ARGV[1] = now (ms timestamp)
-- ARGV[2] = windowMs (duration)
-- ARGV[3] = limit (max requests)
-- ARGV[4] = requestId (unique)
-- ARGV[5] = windowSec (TTL)

-- Remove entries older than window
redis.call('ZREMRANGEBYSCORE', key, 0, now - windowMs)
local current = redis.call('ZCARD', key)

if current < limit then
  redis.call('ZADD', key, now, requestId)      -- Add current request
  redis.call('EXPIRE', key, windowSec + 1)     -- Set expiry
  return {current + 1, 0}
else
  -- Compute Retry-After header
  local oldest = redis.call('ZRANGE', key, 0, 0, 'WITHSCORES')
  return {current, retryAfterMs}
end

Rate Limit Headers Returned:

X-RateLimit-Limit: 10
X-RateLimit-Remaining: 5
X-RateLimit-Reset: 1640000000
Retry-After: 12

---

5. PRISMA SCHEMA - PROPERTY & LISTING MODELS

Property Model

model Property {
  id                   String          @id @default(cuid())
  addressNormalized    String?
  propertyType         PropertyType    // APARTMENT, HOUSE, LAND, COMMERCIAL, etc.
  status               PropertyStatus  // ACTIVE, SOLD, RENTED, REMOVED
  areaM2              Float?
  usableAreaM2        Float?
  bedrooms            Int?
  bathrooms           Int?
  floors              Int?
  floor               Int?
  totalFloors         Int?
  direction           Direction?
  yearBuilt           Int?
  legalStatus         String?
  amenities           Json?
  nearbyPOIs          Json?
  metroDistanceM      Float?
  projectName         String?
  projectDevelopmentId String?
  projectDevelopment  ProjectDevelopment? @relation(...)
  furnishing          Furnishing?
  propertyCondition   PropertyCondition?
  balconyDirection    Direction?
  maintenanceFeeVND   BigInt?
  parkingSlots        Int?
  viewType            String[]        @default([])
  petFriendly         Boolean?
  suitableFor         String[]        @default([])
  whyThisLocation     String?         @db.Text
  createdAt           DateTime        @default(now())
  updatedAt           DateTime        @updatedAt

  listings            Listing[]
  valuations          Valuation[]
  media               PropertyMedia[]

  // Indexes for analytics queries
  @@index([propertyType])
  @@index([district, city])
  @@index([location], type: Gist)                       // PostGIS spatial index
  @@index([district, propertyType])
  @@index([district, city, propertyType])
}

Listing Model (with Analytics Fields)

model Listing {
  id               String          @id @default(cuid())
  propertyId       String
  property         Property        @relation(...)
  agentId          String?
  agent            Agent?          @relation(...)
  sellerId         String
  seller           User            @relation(...)
  transactionType  TransactionType // BUY_SELL, RENT
  status           ListingStatus   @default(DRAFT)
  priceVND         BigInt                              // CHECK: > 0
  pricePerM2       Float?          // Derived for analytics
  rentPriceMonthly BigInt?
  commissionPct    Float?          @default(2.0)
  
  // AI Valuation fields
  aiPriceEstimate  BigInt?         // AVM estimate
  aiConfidence     Float?
  moderationScore  Float?
  moderationNotes  String?
  
  // Analytics tracking
  viewCount        Int             @default(0)
  saveCount        Int             @default(0)
  inquiryCount     Int             @default(0)
  
  // Lifecycle
  featuredUntil    DateTime?
  expiresAt        DateTime?
  publishedAt      DateTime?
  createdAt        DateTime        @default(now())
  updatedAt        DateTime        @updatedAt

  transactions     Transaction[]
  inquiries        Inquiry[]
  orders           Order[]
  priceHistories   PriceHistory[]
  savedByUsers     SavedListing[]

  @@index([status])
  @@index([transactionType])
  @@index([priceVND])
  @@index([sellerId])
  @@index([propertyId])
  @@index([publishedAt])
  @@index([createdAt])
  @@index([status, createdAt(sort: Desc)])             // For market analytics
  @@index([status, transactionType, priceVND])         // For price analysis
}

Price History Model

model PriceHistory {
  id        String   @id @default(cuid())
  listingId String
  listing   Listing  @relation(...)
  oldPrice  BigInt   // CHECK: > 0
  newPrice  BigInt   // CHECK: > 0
  source    String   @default("manual_update")
  changedAt DateTime @default(now())

  @@index([listingId, changedAt(sort: Desc)])  // For trend queries
}

Market Index Model (Pre-calculated Analytics)

model MarketIndex {
  id              String          @id @default(cuid())
  district        String
  city            String
  propertyType    PropertyType
  period          String          // "2024-Q1" or "2024-04" format
  medianPrice     BigInt
  avgPriceM2      Float
  totalListings   Int
  daysOnMarket    Int
  inventoryLevel  Int
  absorptionRate  Float?
  yoyChange       Float?          // Year-over-year % change
  createdAt       DateTime        @default(now())
  updatedAt       DateTime        @updatedAt

  @@unique([district, city, propertyType, period])
  @@index([city, period])
  @@index([period, propertyType])
}

Valuation Model

model Valuation {
  id              String          @id @default(cuid())
  propertyId      String
  property        Property        @relation(...)
  estimatedPrice  BigInt
  confidence      Float           // 0-1 confidence score
  drivers         Json?           // Key price drivers
  comparables     Json?           // Similar properties used
  explanation     String?
  model           String          // "v1" or "v2"
  createdAt       DateTime        @default(now())

  @@index([propertyId, createdAt(sort: Desc)])
  @@index([createdAt])
}

---

6. SHARED GUARDS & DECORATORS

File Locations

apps/api/src/modules/shared/
├── infrastructure/
│   ├── decorators/
│   │   ├── endpoint-rate-limit.decorator.ts      ← Rate limit config
│   │   ├── user-rate-limit.decorator.ts
│   │   └── cacheable.decorator.ts
│   ├── guards/
│   │   ├── endpoint-rate-limit.guard.ts          ← Sliding-window enforcement
│   │   ├── user-rate-limit.guard.ts
│   │   └── feature-listing-throttler.guard.ts
│   ├── middleware/
│   │   ├── correlation-id.middleware.ts          ← Trace ID injection
│   │   ├── csrf.middleware.ts
│   │   ├── request-logging.middleware.ts         ← Audit logging
│   │   └── sanitize-input.middleware.ts
│   ├── cache.service.ts                          ← Cache-aside + Redis
│   ├── redis.service.ts                          ← Redis connection pool
│   ├── logger.service.ts                         ← Structured logging
│   ├── prisma.service.ts
│   ├── field-encryption.service.ts
│   └── filters/
│       └── global-exception.filter.ts            ← Error response standardization
├── domain/
│   ├── domain-exception.ts                       ← Base error class
│   ├── error-codes.ts
│   ├── base-entity.ts
│   ├── domain-event.ts
│   ├── aggregate-root.ts
│   └── value-object.ts
└── utils/
    └── ...

Key Decorators & Guards

@EndpointRateLimit Decorator:

interface EndpointRateLimitOptions {
  limit: number;           // Max requests
  windowSeconds?: number;  // Default 60
  keyStrategy?: 'ip' | 'user';  // Default 'ip'
  adminBypass?: boolean;   // Admins skip limit (default true)
}

// Usage:
@EndpointRateLimit({ limit: 10, windowSeconds: 60, keyStrategy: 'user' })
@UseGuards(EndpointRateLimitGuard, JwtAuthGuard)
async method() {}

Response Standardization (via GlobalExceptionFilter):

interface ErrorResponseBody {
  statusCode: number;
  errorCode: ErrorCode;      // Enum: NOT_FOUND, VALIDATION_FAILED, etc.
  message: string;
  details?: Record<string, unknown>;
  correlationId?: string;    // From CorrelationIdMiddleware
  timestamp: string;
}

Exception Hierarchy

class DomainException extends HttpException {
  constructor(
    errorCode: ErrorCode,
    message: string,
    statusCode?: HttpStatus,
    details?: Record
  )
}

// Specialized exceptions:
class NotFoundException extends DomainException
class ValidationException extends DomainException
class ConflictException extends DomainException
class UnauthorizedException extends DomainException
class ForbiddenException extends DomainException

---

7. DTO PATTERNS IN ANALYTICS

Request DTOs (Query Parameters)

// GET /analytics/market-report?city=...&period=...&propertyType=...
export class GetMarketReportDto {
  @IsString() city: string;
  @IsString() period: string;
  @IsEnum(PropertyType) propertyType?: PropertyType;
}

// POST /analytics/valuation with body
export class PredictValuationDto {
  @IsEnum(PropertyType) propertyType: PropertyType;
  @IsNumber() area: number;
  @IsString() district: string;
  @IsString() city: string;
  @IsNumber() latitude?: number;
  @IsNumber() longitude?: number;
  @IsNumber() bedrooms?: number;
  @IsNumber() bathrooms?: number;
  @IsBoolean() hasElevator?: boolean;
  @IsBoolean() hasParking?: boolean;
  @IsBoolean() hasPool?: boolean;
  @IsNumber() distanceToHospitalKm?: number;
  @IsNumber() distanceToParkKm?: number;
  @IsNumber() distanceToMallKm?: number;
  @IsString() floodZoneRisk?: string;
}

Response DTOs (Handler Return Types)

// From handler
export interface PriceTrendDto {
  district: string;
  city: string;
  propertyType: string;
  trend: PriceTrendPoint[];
}

export interface PriceTrendPoint {
  period: string;
  medianPrice: string;        // Stringified BigInt
  avgPriceM2: number;
  totalListings: number;
}

// From repository
export interface MarketReportResult {
  district: string;
  city: string;
  propertyType: PropertyType;
  period: string;
  medianPrice: string;        // Stringified for JSON safety
  avgPriceM2: number;
  totalListings: number;
  daysOnMarket: number;
  inventoryLevel: number;
  absorptionRate: number | null;
  yoyChange: number | null;
}

---

8. DEPENDENCY INJECTION & MODULE EXPORTS

Analytics Module Registration

@Module({
  imports: [CqrsModule, ListingsModule, AdminModule, ProjectsModule],
  controllers: [AnalyticsController, AvmController],
  providers: [
    { provide: AI_SERVICE_CLIENT, useClass: AiServiceClient },
    { provide: MARKET_INDEX_REPOSITORY, useClass: PrismaMarketIndexRepository },
    { provide: VALUATION_REPOSITORY, useClass: PrismaValuationRepository },
    PrismaAVMService,
    { provide: AVM_SERVICE, useClass: HttpAVMService },  // HTTP → Python AI
    PrismaNeighborhoodScoreService,
    { provide: NEIGHBORHOOD_SCORE_SERVICE, useClass: HttpNeighborhoodScoreService },
    MarketIndexCronService,
    ...CommandHandlers,
    ...QueryHandlers,
    ...EventHandlers,
  ],
  exports: [
    MARKET_INDEX_REPOSITORY,
    VALUATION_REPOSITORY,
    AVM_SERVICE,
    AI_SERVICE_CLIENT,
  ],
})
export class AnalyticsModule {}

Shared Module (Global)

@Global()
@Module({
  imports: [ConfigModule.forRoot(...), EventEmitterModule.forRoot(), PrometheusModule.register(...)],
  providers: [
    LoggerService,
    PrismaService,
    RedisService,        // Redis connection pool
    CacheService,        // Cache-aside pattern + metrics
    EventBusService,
    FieldEncryptionService,
    // Prometheus metrics for cache
    makeCounterProvider({ name: CACHE_HIT_TOTAL, labelNames: ['resource'] }),
    makeCounterProvider({ name: CACHE_MISS_TOTAL, labelNames: ['resource'] }),
    makeCounterProvider({ name: CACHE_DEGRADATION_TOTAL, labelNames: ['resource', 'operation'] }),
    { provide: APP_FILTER, useClass: GlobalExceptionFilter },
  ],
  exports: [PrismaService, RedisService, CacheService, LoggerService, ...],
})
export class SharedModule implements NestModule {
  // Middleware registration:
  configure(consumer: MiddlewareConsumer) {
    consumer
      .apply(CorrelationIdMiddleware, SanitizeInputMiddleware, RequestLoggingMiddleware)
      .forRoutes('*');
    consumer
      .apply(CsrfMiddleware)
      .exclude([{ path: 'auth/login', method: RequestMethod.POST }, ...])
      .forRoutes('*');
  }
}

---

9. KEY PATTERNS & CONVENTIONS

Cache Key Building

CacheService.buildKey(CachePrefix.MARKET_TREND, district, city, propertyType, periods)
// → "cache:market:trend:{district}:{city}:{propertyType}:{periods}"

Error Handling Pattern

async execute(query: Query): Promise<Result> {
  try {
    const cacheKey = CacheService.buildKey(...);
    return this.cache.getOrSet(cacheKey, async () => {
      // Business logic
    }, CacheTTL.MARKET_DATA, 'metric_label');
  } catch (error) {
    if (error instanceof DomainException) throw error;  // Re-throw domain errors
    this.logger.error(`Failed to ...`, error.stack, this.constructor.name);
    throw new InternalServerErrorException('User-facing message');
  }
}

Repository Pattern (Dependency Inversion)

// domain/repositories/market-index.repository.ts
export const MARKET_INDEX_REPOSITORY = Symbol('MARKET_INDEX_REPOSITORY');
export interface IMarketIndexRepository { ... }

// infrastructure/repositories/prisma-market-index.repository.ts
@Injectable()
export class PrismaMarketIndexRepository implements IMarketIndexRepository { ... }

// In handlers:
@Inject(MARKET_INDEX_REPOSITORY) private readonly repo: IMarketIndexRepository

Quota & Rate Limit Stack

@ApiBearerAuth('JWT')
@EndpointRateLimit({ limit: 10, windowSeconds: 60, keyStrategy: 'user' })
@UseGuards(EndpointRateLimitGuard, JwtAuthGuard, QuotaGuard)  // Order matters!
@RequireQuota('analytics_queries')  // Quota resource name

Query Handler Naming

• Query class: GetPriceTrendQuery
• Handler: GetPriceTrendHandler (decorated with @QueryHandler(GetPriceTrendQuery))
• DTO: GetPriceTrendDto (response type)
• Query class file: get-price-trend.query.ts
• Handler file: get-price-trend.handler.ts

---

10. QUICK REFERENCE: PATHS & KEYS

Module Root:

apps/api/src/modules/analytics/

Controllers:

analytics.controller.ts    → /analytics/*
avm.controller.ts          → /avm/*

Cache Prefixes (for market analytics):

cache:market:report       → Market report data
cache:market:trend        → Price trends
cache:market:heatmap      → Heatmap data
cache:market:district     → District statistics
cache:valuation           → Property valuations

Shared Module Exports:

PrismaService             → Database
RedisService              → Redis connection
CacheService              → Cache-aside + metrics
LoggerService             → Structured logging
EventBusService           → Event emission

Key Decorators:

@EndpointRateLimit({...}) → Per-endpoint sliding-window rate limit
@RequireQuota('...')       → Subscription quota check
@UseGuards(...)            → Auth, rate limit, quota guards

TTLs:

MARKET_DATA: 1800         → 30 minutes (price trends, historical)
MARKET_REPORT: 900        → 15 minutes (report summaries)
HEATMAP: 300              → 5 minutes (heatmap tiles)
DISTRICT_STATS: 300       → 5 minutes (statistics)