This project focuses on identifying why customers are leaving a telecommunications company (Churn Rate: 26.58%). By analyzing a dataset of 7,043 customers, I identified key behavioral patterns and service gaps. This repository contains the full end-to-end pipeline: from raw data cleaning in Python to an interactive Power BI dashboard.
- Python (Pandas, Seaborn, Matplotlib): Data auditing, cleaning, and exploratory data analysis (EDA).
- Microsoft Excel: Feature engineering and pivot table summaries.
- Power BI: Interactive data visualization and DAX measures.
- Gamma AI: Executive presentation for stakeholders.
Data/: Contains the raw CSV and the final formatted Excel datasets.Python Sciprts/: Jupyter Notebook containing the data cleaning and correlation analysis.Visualizations/: Dashboard screenshots and individual charts showing key drivers of churn.Reports/: Full project documentation and technical notes.Presentation/: Executive-level slides (PDF/PPTX) summarizing the business impact.
- The Support Gap: 63% of churned internet users did not have Technical Support. Providing support is the #1 way to increase loyalty.
- Fiber Optic Risk: Despite being a premium service, Fiber Optic has a 41.89% churn rate, likely due to high costs ($74.40 avg) without bundled support.
- Contract Maturity: Customers on Month-to-Month contracts represent the highest attrition risk. Long-term (1-2 year) contracts significantly stabilize revenue.
The dashboard allows users to filter by Payment Method, Internet Service, and Tenure to identify "at-risk" customer segments in real-time.
- Contract Migration: Offer a 10% discount to transition high-risk Month-to-Month users to 1-year plans.
- Support Bundling: Include "Premium Tech Support" for all Fiber Optic plans to reduce service frustration.
- Retention Milestones: Implement automated check-ins for new customers at the 3-month and 6-month marks, as these are the peak churn windows.
- Read the
Reports/Telco customer analysis documentation.docxfor the full methodology. - Open
Python Sciprts/analysis.ipynbto see the data cleaning code. - View
Presentation/Telco-Customer-Churn-Analysis.pdffor the executive summary.
