What does the Deep Check mode actually do?

Deep Check mode pulls raw source data and manually recalculates statistical averages to compare them against the cached values in your database, identifying discrepancies larger than 2%.

What is the Data Cascade Problem in this context?

The Data Cascade Problem occurs when missing data on a specific date corrupts rolling averages (like L5 or L10) and ML features for up to 21 days forward, leading to degraded prediction accuracy.

How does the skill help fix missing data?

It provides a specific remediation plan in the correct dependency order, typically starting with Phase 3 Analytics regeneration followed by Phase 4 Precompute updates.

Can I validate data for a specific NBA player?

Yes, by using the --player flag followed by a name, the skill performs a deep dive into that specific player's game history and verifies their rolling average integrity.

Historical Data Integrity Validator

Name: Historical Data Integrity Validator
Author: najicham

bynajicham

0•

Ciencia de Datos y ML

Audits historical data completeness and quality across date ranges to prevent downstream prediction corruption and feature degradation.

The Historical Data Integrity Validator is a specialized tool for the NBA stats scraper ecosystem designed to solve the 'Data Cascade Problem'—where a single day of missing data can corrupt rolling averages and ML features for up to 21 days. It provides deep visibility into BigQuery data layers, identifying gaps in raw, analytics, and prediction tables while offering specific remediation steps. Whether performing a routine weekly health check or investigating specific statistical anomalies, this skill ensures that the data powering your models remains accurate, consistent, and fully backfilled in the correct dependency order.

Características Principales

01Interactive guided validation for custom date ranges and player-specific audits

02Cascade impact assessment to identify affected future rolling averages

03Automated gap detection across raw, analytics, and prediction data layers

040 GitHub stars

05Deep check mode for verifying cached stats against manual source recalculations

06Step-by-step remediation plans with ordered commands for pipeline recovery

Casos de Uso

01Verifying the success of backfill operations after a scraper failure

02Investigating ML model performance degradation by auditing feature store quality trends

03Performing weekly pipeline health checks to ensure 100% data coverage

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add najicham/nba-stats-scraper validate-historical

For use in Claude.ai and ChatGPT

Download Skill