Skip to content

Complaint analysis for a major US cable provider shows a 220% spike in June, mainly on data caps and internet issues. Using Python, Pandas, BERTopic, and NLP, we aim to cut complaints by 50%, resolve 95% by Q4, and uncover top 3 root causes within 2 weeks.

Notifications You must be signed in to change notification settings

ranggaakhli/customer_complaints

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 

Repository files navigation

📞 Customer Complaint Analytics – Cable & Internet Service Provider (US)

🎯 Business Objective (SMART)

  • Specific: Reduce repeat complaints and boost customer satisfaction.
  • Measurable: Cut complaint volume by 50% in 6 months; reach 90% resolution rate by Q3, 95% by Q4.
  • Achievable: Identify top 3 root causes of recurring issues.
  • Relevant: Complaints drive churn and hurt brand reputation.
  • Time-bound: Deliver insights in 2 weeks to inform Q3–Q4 strategy.

🧭 Analysis Outline

  1. Executive Summary
  2. Dataset Overview & Cleaning
  3. NLP with BERTopic: Identifying Core Complaint Themes
  4. Complaint Volume Trends (Monthly)
  5. Source Channel Analysis
  6. Geographic Breakdown (State & City)
  7. Complaint Status Performance
  8. Topic Label Breakdown
  9. “Filed on Behalf” Behavior
  10. Pareto Analysis
  11. WordCloud
  12. Conclusion & Recommendations

📌 Executive Summary

  • 📈 +220% complaint surge in June, mainly around data caps and internet services
  • 🔥 Top keywords: "Disney", "internet", "services"
  • ✅ Resolution rate: 76.75%, but 520 cases remain unresolved
  • 🌎 Top complaint states: Georgia, Florida, California
  • 🏙️ Highest city-level complaints: Atlanta (with high unresolved rate)
  • Chicago shows strong complaint resolution—potential benchmark city
  • ⏱️ SLA: 53.32% unresolved cases are <7 days old; 46.7% are older
  • 🎯 Pareto Insight: Focus on data cap and internet service issues for maximum impact

About

Complaint analysis for a major US cable provider shows a 220% spike in June, mainly on data caps and internet issues. Using Python, Pandas, BERTopic, and NLP, we aim to cut complaints by 50%, resolve 95% by Q4, and uncover top 3 root causes within 2 weeks.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published