nreddie7702

Churnguard Analytics — Data skill for Claude Code

Data community

End-to-end SaaS revenue leakage & churn prevention analytics platform — synthetic data generation, SQL analysis, ML churn prediction (ROC-AUC 0.958), Claude-powered AI Analyst & Agent, and a 5-page Po.

How to install Churnguard Analytics

This entry records only its repository, not the path inside it, so there is no exact command to give. Open nreddie7702/churnguard-analytics and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Churnguard Analytics does

Alternatives in Data

Context Mode — Benchmark Results — Benchmarked against real outputs from popular Claude Code MCP servers, Skills, and dev tools 5.6k ★
MCP Server Chart — 🤖 A visualization mcp & skills contains 25+ visual charts using @antvis 4k ★
Arcgis To Portaljs — Migrate a whole ArcGIS Hub site (opendata.arcgis.com or a Hub Premium custom domain) into a PortalJS Arc porta 2.3k ★

README

ChurnGuard Analytics

**An end-to-end SaaS revenue leakage & churn prevention analytics platform** — from synthetic data generation through SQL analysis, machine learning, GenAI-powered querying, and an executive-ready Power BI dashboard.

Built as a portfolio project to demonstrate the full data analyst workflow on a realistic SaaS dataset: 1,200 customers, 2 years of subscription/invoice/usage/support activity, and a central business question — **why are we losing revenue, and which customers are at risk?**

Key Findings

Metric	Value
Total Active MRR	$320.41K
Active Customers	632
Logo Churn Rate	47%
Churned Customers	566
Average Revenue Per Customer (ARPU)	$291
Churn Model Accuracy (ROC-AUC)	0.958
Customers Flagged High Churn Risk	127 (20%)
Open Support Tickets	294
Failed Payments	526 ($109.64K)

**Headline insight:** Enterprise-tier customers drive the most revenue *and* are the most loyal (lowest churn risk, ~0.09 average), while Basic-tier customers are both the least profitable *and* the most likely to churn (~0.29 average risk) — a classic SaaS pattern that directly informs where retention effort should be focused.

What's Inside

Synthetic data generation — 1,200 SaaS customers, 2 years of activity, calibrated to realistic SaaS benchmarks (not random noise)
Data cleaning — deduplication, inconsistent formatting fixes, referential integrity checks
25 SQL business questions — window functions, CTEs, cohort analysis
EDA & feature engineering — statistical testing, correlation analysis
Machine learning — Random Forest churn classifier, ROC-AUC 0.958
GenAI layer (Claude-powered):
- AI Analyst — natural language Q&A grounded in real SQL query results (never hallucinates numbers)
- AI Agent — autonomous, multi-step investigation using both SQL and a sandboxed Python analysis tool
Power BI dashboard — 5 interactive pages: Exe