📊 Customer Response Model for Direct Marketing Campaigns

Project Overview:

This project aimed to build a response model for a marketing department in a retail food company. The goal was to predict which customers were more likely to respond positively to a new product offer, based on data from a previous marketing campaign. By identifying these customers, the company can optimize its direct marketing campaigns, leading to improved profitability.

This project was completed as part of a group assignment for the Programming for Data Science course.

The project involved various steps of data exploration, feature extraction, and analysis, culminating in a model that can efficiently predict future responses to marketing offers.

---

🎯 Project Objective:

The main objective of this project is to develop a model that can predict which customers are likely to respond positively to a direct marketing campaign. By analyzing data from a previous campaign, the goal was to:

• Discover patterns, trends, and correlations within the customer data. 📈
• Create a set of discriminating characteristics that define the customers who are more likely to purchase the new product. 🛍️
• Build a model that can generalize these findings to the entire customer base, ensuring that future campaigns are more profitable. 💡

---

📊 Data Description:

The dataset consists of 2,240 customers who were contacted as part of a pilot marketing campaign. It includes 28 socio-demographic, firmographic, and behavioral features, as well as a flag indicating whether the customer responded to the campaign by purchasing the product. The features are categorized into six main groups:

1. Socio-Demographic: Includes variables like BirthYear, Education, MaritalStatus, etc. 👤
2. Frequency: Data about the customer's purchase and interaction frequency, e.g., NDealsPurchases, NWebVisitsMonth, etc. 🔄
3. Monetary Value: Information on spending patterns, e.g., MntWines, MntMeatProducts, etc. 💰
4. Customer Lifetime and Satisfaction: Customer-specific data like CustomerFrom, Complain. 🕒
5. Campaign Response: Five response flags corresponding to different campaigns. 📞
6. Ultimate Campaign Cost and Revenue: Data related to the cost and revenue generated from the pilot campaign. 📉📈

---

📝 Guidelines:

The project consists of a series of steps that will be completed through Jupyter Notebooks and a spreadsheet. The files and their purposes are as follows:

1. combine.ipynb: Combines the separate data files from the six extraction calls. 🔗
2. preprocess.ipynb: Describes the structure of the dataset, handles missing values, duplicates, and feature transformation. 🧹
3. extract.ipynb: Extracts useful features and evaluates feature correlation and redundancy. 🔍
4. explore1.ipynb: Assesses individual features' ability to discriminate between respondents and non-respondents, using Scaled mean deviation, Spearman vs Pearson correlation and the chi^2 test of independence. 🔑
5. explore2.ipynb: Explores the joint discrimination power of features, aiming to extract precise and representative characteristics of respondents. 🧠
6. conclusion.ipynb: Summarizes the findings and provides a detailed yet concise description of the key discriminating features. 📝
7. respondents.xlsx: Contains the subset of customers who are more likely to respond to future campaigns based on the selected discriminating features. 📊

---

The estimated profit was 421MU.

---

🚀 Installation and Setup:

1. Clone the repository:

   git clone https://github.com/MGN19/Customer-Response-Model-in-Direct-Marketing.git

▶️ How to Run: All project tasks were conducted in JupyterLab. Run each notebook in the given order:

1. combine.ipynb
2. preprocess.ipynb
3. extract.ipynb
4. explore1.ipynb
5. explore2.ipynb
6. conclusion.ipynb

Excel Output: After completing the notebooks, the final list of respondents were saved as respondents.xlsx, which can be used for the next campaign. 📊