📊 Megaline Telecom: Customer Behavior & Revenue Optimization Analysis

Data Analysis • A/B Testing • Statistical Evaluation • Python • Pandas • Seaborn

---

🚀 Key Results (Quick Snapshot)

• 📈 Ultimate plan generates higher average monthly revenue per user
• 📊 Statistically significant difference (p-value = 3.63e-16, α = 0.01)
• 🌐 Internet overage is the #1 revenue driver
• 💡 Identified high-value segment for plan upgrades → revenue growth opportunity

---

🔍 Project Summary

Megaline, a telecom company, needed to determine which of its two pricing plans generates higher revenue and how customer behavior impacts profitability.

This project analyzes user activity and applies statistical testing to identify revenue drivers and inform data driven pricing strategy decisions.

---

📌 Project Highlights

• Merged and transformed multiple datasets into a unified user level dataset

• Engineered features to calculate monthly revenue based on pricing rules and overage charges

• Conducted exploratory data analysis to uncover behavioral differences between customer segments

• Performed statistical hypothesis testing (two-sample t-test, p < 0.01) to validate revenue differences

• Identified key revenue drivers, including internet usage and overage frequency

• Developed actionable business recommendations to improve revenue and customer retention

---

🎯 Objective

• Compare customer usage patterns across Surf and Ultimate
• Calculate monthly revenue per user
• Analyze excessive usage (calls, texts, data) and overage charges
• Statistically test whether one plan earns more than the other
• Provide clear business recommendations to Megaline

---

🛠️ Tools & Skills

• Python (Pandas, NumPy)
• Data Visualization (Matplotlib, Seaborn)
• Statistical Testing (SciPy, t-test)
• Feature Engineering
• Business Analytics & Revenue Modeling
• Hypothesis Testing & A/B Testing

---

🚀 Project Workflow

1️⃣ Data Preparation

• Loaded multiple datasets: calls, messages, internet, and user info
• Merged them into a single monthly user level dataset
• Handled missing values and corrected data types
• Calculated monthly totals for call minutes, number of messages, and internet usage (MB/GB)

2️⃣ Exploratory Data Analysis (EDA)

• Analyzed how much users call, text, and use data on each plan
• Visualized distributions and usage patterns
• Compared Surf vs. Ultimate users on:
  • Call duration
  • Number of SMS
  • Internet usage

3️⃣ Revenue Modeling

For each user and month, I computed total revenue based on the tariff rules:

• Fixed monthly fee
• Included limits for minutes, messages, and data
• Overage charges when users exceeded their limits

This produced monthly revenue per user, which became the main metric for comparison.

---

🧪 Statistical Hypothesis Testing

• H₀ (null): There is no difference in average monthly revenue between Surf and Ultimate plans
• H₁ (alternative): The average monthly revenue differs between the plans
• Test: Two sided t-test
• Significance level: α = 0.01

The test returned:

• p-value = 3.63 × 10⁻¹⁶

Since p ≪ 0.01, the result is highly significant.

---

📈📊 Key Insights

• Internet usage is the primary driver of revenue, contributing the largest share of overage charges across both plans.

• Surf users consistently exceed plan limits more frequently, leading to unpredictable billing and higher variability in monthly revenue.

• Ultimate users generate more stable and predictable revenue due to higher base pricing and lower overage dependency.

• High usage customer segments represent the most valuable revenue opportunity, particularly when transitioned to higher tier plans.

• The revenue distribution between plans shows clear separation, indicating fundamentally different user behaviors and monetization patterns.

• Statistical testing confirmed a significant difference in average revenue between plans (p < 0.01), validating that observed trends are not due to random variation.

---

💰 Revenue Strategy Insights

• Overage-based revenue (Surf plan) is less predictable but can generate higher short term gains from heavy users

• Subscription based revenue (Ultimate plan) provides consistent and reliable income with lower volatility

• Customers exceeding limits frequently are strong candidates for plan upgrades, representing a key upsell opportunity

• Pricing structure directly influences customer behavior, with lower-tier plans encouraging overage usage patterns

---

🏁 Final Results & Business Impact

• Higher-earning plan:

  • Ultimate appears to be more profitable overall because of its higher base fee and stable revenue per user

• Statistical Significance:

  • p-value = 3.63 × 10⁻¹⁶ < 0.01 → Reject H₀
  • The revenue difference between plans is not due to chance

• Business Recommendations:

  • Encourage heavy internet users to upgrade to Ultimate
  • Offer targeted add on data packages for Surf users who frequently exceed limits
  • Use usage patterns to segment customers and personalize offers
  • Consider marketing campaigns that move high usage Surf customers to Ultimate to stabilize and grow revenue

---

📊 Key Visualizations

Ultimate plan users exhibit higher median data usage and greater variability, indicating heavier and less predictable consumption patterns compared to Surf users.

Ultimate users also demonstrate higher call volume, suggesting increased engagement and usage intensity.

📦 Repository Structure

Megaline-Telecom-Analysis/ ├── data/ │ ├── calls.csv │ ├── messages.csv │ ├── internet.csv │ └── users.csv ├── notebooks/ │ └── megaline_analysis.ipynb ├── images/ │ ├── internet_usage.png │ └── call_minutes.png └── README.md

✅ Conclusion

This analysis revealed clear and statistically significant differences in revenue between Surf and Ultimate plans (p < 0.01).

Ultimate users generate more stable and predictable revenue, while Surf users contribute higher variability due to frequent overage charges.

From a business perspective, the most valuable opportunity lies in identifying high-usage Surf customers and strategically converting them to Ultimate plans.

This project demonstrates the ability to combine data analysis, statistical validation, and business strategy to drive data-informed decisions.

Key outcomes:

• Identified usage patterns across customer segments to determine which plans generate the highest revenue
• Applied statistical hypothesis testing to validate differences in plan performance
• Uncovered insights showing how high usage customers can be strategically targeted for plan upgrades

Business Impact:

• Recommended transitioning high usage “Surf” users to the “Ultimate” plan to increase revenue
• Suggested targeted add on packages to reduce churn and improve customer retention
• Demonstrated how data driven segmentation can support personalized marketing strategies

This project highlights my ability to:

• Perform end-to-end exploratory data analysis using Python
• Apply statistical methods to real world business problems
• Translate data insights into actionable business decisions

I’m excited to apply these skills in a Data Analyst or Data Scientist role to help organizations make smarter, data driven decisions.

---

⚙️ How to Run

1. Clone the repository
2. Open the notebook in Jupyter
3. Run all cells to reproduce the analysis

Dependencies:

• Python 3.x
• pandas, numpy
• matplotlib, seaborn
• scipy