Bank Marketing Campaign Prediction

Authors: Hala Arar, Fazeeia Mohammad, Rong Wan

Project Summary

This project aims to enhance bank marketing campaigns by using machine learning to predict whether a customer will subscribe to a term deposit based on demographic and campaign-related data. The goal is to develop a model that can more effectively target potential customers, improving resource allocation and reducing marketing costs by excluding unlikely subscribers.

The project explores several machine learning models, including Logistic Regression and Decision Trees. These models are trained on customer data, with preprocessing techniques like feature scaling and one-hot encoding applied to prepare the data for analysis. The outcome of this work demonstrates how banks can leverage machine learning to implement more effective, data-driven marketing strategies, which can lead to better customer acquisition and optimized campaign performance.

The Logistic Regression model achieved an accuracy of 88.5%, with a focus on minimizing false positives, resulting in a precision of 0.70 and recall of 0.20. Despite its high precision, the model's low recall indicates that it misses a significant portion of actual subscribers. The Decision Tree model, with an accuracy of 89.7%, demonstrated better recall (0.23) but at the cost of increased false positives (126). Both models highlight the class imbalance in the dataset, where non-subscribers are far more prevalent than subscribers. The Logistic Regression model is more suitable when minimizing false positives is prioritized, whereas the Decision Tree model is more effective in identifying potential subscribers but may require further regularization to reduce overfitting.

Strategic recommendations for targeted marketing, personalized offers, and campaign timing are proposed to optimize resource allocation and improve conversion rates. Future model iterations should focus on improving both precision and recall to enhance marketing efforts and increase return on investment.

Report

The final report can be found here.

How to Run the Analysis

Option 1: Run Using Docker

1. Install Docker

Download and install Docker Desktop for your operating system and ensure it is running.

2. Clone the repository

Open your terminal and run the following command to clone the project repository to your local machine:

git clone https://github.com/mindy001/BankMarketingPreditions-.git

Navigate to the project directory

cd BankMarketingPreditions-

3. Run the Docker Container Using Docker Compose

Run the following command in your terminal to pull the project’s Docker image:

docker-compose up

4. Access JupyterLab

You will see a URL in the terminal. Open the link in your browser to access the JupyterLab environment.

Option 2: Run Using the Make file

1. Clone the repository

git clone https://github.com/mindy001/Group37DSCI522.git

Navigate to the project directory

cd BankMarketingPreditions-

2. Set up the environment

conda env create -f env/environment.yml conda activate bankenv

3. Run the Report

To generate the report, run the following command in your terminal:

make all

4. Open the Report

The final report is available as a PDF. You can view the completed analysis by opening the bank_marketing_analysis.pdf file.

Option 3: Run Locally

1. Clone the repository

git clone https://github.com/mindy001/Group37DSCI522.git

2. Set up the environment

conda env create -f env/environment.yml conda activate bankenv

3. Run the Analysis

After activating the environment, you can run the analysis script or Jupyter notebook called bank_marketing_analysis.ipynb

4. Open the Report

The final report is available as a PDF. You can view the completed analysis by opening the bank_marketing_analysis.pdf file.

Running Project scripts

1. Change your directory to the current project directory using the cd command from bash.

2. Scripts are run using the click command in the root of the project. More details about the scripts can be found in the scr directory.

3. These are command lines to run the python files: python 01_download.py --directory data/bankmarketing/bank-additional/bank-additional/ --filename bank-additional-full.csv

    python 02_clean_data.py --input_path data/bankmarketing/bank-additional/bank-additional/bank-additional-full.csv --output_path data/cleaned_bank_data.csv
   
    python 03_explory_analysis.py --cleaned_data_path data/cleaned_bank_data.csv --output_prefix results/eda
   
    python 04_model_LR.py --input_path ./cleaned_data.csv --model_output_path ./model/logistic_regression_model.pkl --confusion_matrix_output .results/eda/confusion_matrixLR.png
   
    python 04_model_DT.py --input_path ./cleaned_data.csv --model_output_path ./model/decision_tree_model.pkl --confusion_matrix_output .results/eda/confusion_matrixDT.png
   

Dependencies

Docker is a container solution used to manage the software dependencies for this project. The Docker image used for this project is based on the quay.io/jupyter/minimal-notebook:notebook-7.0.6 image. Additional dependencies are specified in the Dockerfile.

To run the analysis and work with the code, you will need to install the following Python packages. These are automatically included in the environment.yml file, but here is the full list for reference:

• Python 3.10
• matplotlib==3.9.2
• numpy==1.23.5 (Compatible with scikit-learn and scipy)
• pandas==1.5.3 (Stable version of pandas)
• scikit-learn==1.2.2 (Includes tools like train_test_split, GridSearchCV, StandardScaler, OneHotEncoder, KNeighborsClassifier, LogisticRegression, DecisionTreeClassifier, etc.)
• seaborn==0.13.2
• altair==5.1.0
• scipy==1.10.1 (Compatible with numpy 1.23.x)
• ipython==8.12.0
• nbformat==5.9.2
• jupyter==1.0.0
• jupyterlab==4.0.5
• quarto==1.3.433
• pip==24.0
• altair-ally==0.1.1 (Installed via pip)
• pandera==0.8.1 (Installed via pip)
• tabulate (Installed via pip)
• ucimlrepo
• click

These dependencies are necessary for data processing, model building, evaluation, and visualization.

License

This project is licensed under CC0 1.0 Universal (Creative Commons Public Domain Dedication). By applying this license, the creator voluntarily waives all copyright and related rights, allowing anyone to use, modify, distribute, or build upon the work for any purpose, including commercial purposes, without the need for permission or attribution.

References

Meshref, H. (2020). Predicting loan approval of bank direct marketing data using ensemble machine learning algorithms. International Journal of Circuits, Systems and Signal Processing, 14, 117. https://doi.org/10.46300/9106.2020.14.117

Moro, S., Rita, P., & Cortez, P. (2014). Bank Marketing [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C5K306

Wang, D. (2020). Research on bank marketing behavior based on machine learning. AIAM2020: Proceedings of the 2nd International Conference on Artificial Intelligence and Advanced Manufacture, 150–154. https://doi.org/10.1145/3421766.3421800

Xie, C., Zhang, J.-L., Zhu, Y., Xiong, B., & Wang, G.-J. (2023). How to improve the success of bank telemarketing? Prediction and interpretability analysis based on machine learning. Computers & Industrial Engineering, 175, 108874. https://doi.org/10.1016/j.cie.2022.108874

Zaki, A. M., Khodadadi, N., Lim, W. H., & Towfek, S. K. (2024). Predictive analytics and machine learning in direct marketing for anticipating bank term deposit subscriptions. American Journal of Business and Operations Research, 11(1), 79-88. https://doi.org/10.54216/AJBOR.110110