The goal of this project is to analyze the customer data of Megaline, a telecommunications company offering two prepaid plans: Surf and Ultimate.
The commercial department seeks to identify which plan generates more revenue to optimize the advertising budget. Using data from 500 customers, the analysis will focus on customer behavior, usage patterns, and revenue generation to make data-driven decisions.
The project consists of the following steps:
1. Plan Usage and Customer Behavior:
- Analyzed call, message, and internet usage of Megaline customers to understand differences in usage patterns between the Surf and Ultimate plans.
- Monthly limits for minutes, messages, and data were compared across plans to determine which plan exceeded limits more often and where additional charges were accrued.
2. Revenue Generation Analysis:
- Calculated the total revenue generated per user by considering plan fees and any additional costs from exceeding limits (calls, messages, data).
- Determined the average monthly revenue for each plan and performed statistical tests to compare the revenue generated by the Surf and Ultimate plans.
- Segmented analysis by region to identify potential differences in revenue generation across different locations.
3. Hypothesis Testing:
- Tested hypotheses to check if the average revenue per user differs significantly between the Surf and Ultimate plans.
- Analyzed whether the average revenue in the NY-NJ area is different from that in other regions.
Data Analysis & Cleaning:
- Utilized Python libraries such as pandas for data manipulation, merging datasets, and handling missing values.
- Ensured data accuracy by converting data types and identifying and correcting data errors.
Exploratory Data Analysis (EDA):
- Visualized customer usage patterns using histograms and descriptive statistics to understand data distributions.
- Calculated metrics like averages, variances, and standard deviations for call duration, message counts, and internet usage.
Revenue Calculation:
- Calculated monthly revenue for each user by accounting for their usage and associated costs based on their plan.
- Created a summary of total revenue per plan to identify which plan contributes more to the company’s profits.
Hypothesis Testing:
- Applied statistical tests to compare average revenues between different user groups and regions.
- Formulated null and alternative hypotheses and selected appropriate significance levels to validate assumptions.
Usage Behavior by Plan:
- Surf Plan: Customers using the Surf plan typically reached their usage limits faster, leading to additional charges for calls, messages, and data. On average, Surf plan users utilized around 400-500 minutes, 30-45 messages, and 12-14 GB of data per month.
- Ultimate Plan: The Ultimate plan customers had more generous usage limits, and only a small proportion exceeded their monthly allowances. They typically used around 1000-1200 minutes, 300-500 messages, and 20-25 GB of data. Revenue Comparisons:
Monthly Revenue:
- The average monthly revenue for Surf users was lower than that for Ultimate users, primarily due to the Ultimate plan's higher base monthly fee.
- However, additional usage charges for Surf plan customers increased their total revenue contribution significantly, especially for those who frequently exceeded plan limits.
Total Revenue Generation:
- Surf Plan: Although the base fee was lower, the total revenue per customer was often inflated by overage charges on calls, texts, and data.
- Ultimate Plan: Higher base revenue due to the plan's cost, with fewer additional charges for overages.
Customer Behavior Insights:
- Peak Usage Periods: Surf users were more likely to accumulate overage charges in the final days of the month, indicating a pattern of exceeding limits as the month progressed.
- Revenue Drivers: For Surf users, overage charges were a key driver of revenue, while for Ultimate users, the base fee accounted for nearly all the revenue, with minimal overage.
Statistical Testing Results:
-
Revenue Hypothesis Testing:
- The statistical analysis confirmed that the average revenue per user between Surf and Ultimate plans was significantly different.
- Ultimate plan users consistently generated more stable revenue, while Surf plan revenue had greater variability due to overage charges.
-
Regional Revenue Differences:
- Users in the NY-NJ area demonstrated slightly higher average revenue compared to other regions, potentially due to different usage patterns or a higher tendency to exceed plan limits.
Plan Recommendations Based on Findings:
- For Surf Users: Increasing plan limits or encouraging users to switch to the Ultimate plan could stabilize revenue and reduce overage charge fluctuations.
- For Ultimate Users: Continuing to market the Ultimate plan as an all-inclusive option helps retain high-value users who need larger allowances without unexpected charges.
- Marketing Focus: Target high-usage Surf users for plan upgrades, especially those who frequently exceed limits.
- Data Cleaning and Preparation: Improved skills in handling multiple datasets, correcting data inconsistencies, and preparing data for analysis.
- Revenue and Cost Analysis: Developed a structured approach to calculating and comparing revenue across different user groups and plans.
- Statistical Testing Competence: Gained experience in hypothesis formulation and statistical testing to make data-backed decisions on revenue and user behavior.
- Data Visualization & Insights: Enhanced ability to use visualizations to identify trends, compare plan performance, and communicate findings effectively.
-
Clone the Repository git clone https://github.com/realdanizilla/Megaline.git
-
Explore the Notebooks and datasets Review the Jupyter Notebooks and datasets for data exploration, preprocessing, and model building.
-
Execute the Analysis Run the provided code to preprocess the data, train classification models, and test their accuracy in recommending plans.
-
Check Model Performance Evaluate model predictions to see how well they recommend the appropriate telecom plan for customers.