Amitmelamed277/marketing-campaign-eda

# 📊 Customer Personality Analysis & Marketing ROI

## 1. Project Overview
This project involves an end-to-end Exploratory Data Analysis (EDA) of a retail marketing dataset containing 2,240 customers and 29 initial features. The primary objective is to analyze customer purchasing behaviors, understand demographics, and identify the "Golden Customer" profile to optimize future marketing campaigns and maximize Return on Investment (ROI).

[](https://www.youtube.com/watch?v=XLrvb6GRse0)

## 2. Data Cleaning & Preprocessing Decisions
Before analyzing the data, a thorough data preparation process was executed to ensure the models and visualizations are built on reliable information.

* **Handling Missing Values:** * *Decision:* 24 missing values were detected in the `Income` column.
* *Action:* Filled these missing values using the **Median** (`51,381`). The median was chosen over the mean to avoid skewness caused by extreme high-income earners.
* **Outlier Detection & Removal:**
* *Age Outliers:* Detected customers with birth years before 1900 (e.g., 120+ years old). These were removed as input errors.
* *Income Outliers:* Identified an extreme outlier with an income of `$666,666`, which distorted our visualizations. Filtered the dataset to include only incomes below `$200,000`.

## 3. Feature Engineering & Dimensionality Reduction
To properly answer our research questions, new logical features were created:
* `Age`: Calculated by subtracting `Year_Birth` from the campaign year (2014).
* `Total_Spent`: Aggregated sum of all product categories (Wines, Fruits, Meat, Fish, Sweets, Gold).
* `Enrollment_Year`: Extracted the specific joining year from the full `Dt_Customer` date.

*Dimensionality Reduction:* Dropped redundant and technical columns (`ID`, `Z_CostContact`, `Z_Revenue`, `Year_Birth`, full `Dt_Customer`) to clean the dataset for future Machine Learning models.

## 4. Exploratory Data Analysis (EDA): Questions & Insights

### Q1: What is the relationship between customer income and spending on Wine?


* **Insight:** There is a strong, clear positive correlation. As annual income increases, the amount spent on wine increases significantly.

### Q2: Does the presence of children in the household increase meat consumption?


* **Insight:** Counter-intuitively, households with **0 children** spend almost twice as much on meat products compared to families with children. This suggests child-free households are purchasing premium, expensive meat cuts, while families buy basic necessities.

### Q3: How does education level impact overall spending power?


* **Insight:** Customers with higher academic degrees (Master's and PhD) not only earn significantly higher incomes but are also the primary drivers of revenue, having the highest `Total_Spent`. Customers with 'Basic' education spend almost nothing.

### Q4: Does customer loyalty (seniority) translate to higher spending?


* **Insight:** Veteran customers who joined the company in 2012 have a significantly higher average spending basket compared to newer customers acquired in 2014. Retaining old customers is highly profitable.

## 5. Final Conclusion
The data clearly defines our **"Golden Customer"**: A mature individual, highly educated (Master's/PhD), high-income earner, without children at home, and a veteran member of our network. By targeting this specific demographic, the marketing department can drastically reduce ad spend waste and increase the overall ROI.