Predicting the Success of Term Deposit Subscriptions in Telemarketing Campaigns
The past few years leading up to the conclusion of 2024, ushered in unprecedented and uncertain economic financial changes in the US’s economy of drastic proportions. America has been mired by a dismal job market, dramatic increases in home prices, groceries, and other common goods, products, and services. However, what's interesting is that this dataset provides details gathered from a Portuguese telemarketing bank strategy conducted entirely by phone during a time when Portugal’s financial and economic systems were facing similar pressures and struggles. In today’s digital age where unsolicited privacy of one’s personal information is highly valued, engaging with customers in this manner is a risky move since most individuals find telemarketing campaigns to be intrusive and annoying. Therefore, the exponential increases in costs across all industries, the slow economy, and sluggish job market are a few reasons peaking our interest in exploring this domain and dataset, serving as a major influence on our team's interest in comparatively investigating the predictive proweress of various statistical models such as logistic regression machine learning tools, LDA, QDA, and Random Forest. The plan for our predictive model is to provide financial professionals with insights in determining which potential customers from various socio-economic backgrounds were more likely to subscribe to banking term deposits through telemarketing campaigns.
- Shanara Hawkins @ShanaraTech
- Karthik Radhakrishnan @kradhakrishnan0714
We collaborated as a group to implement all the major components and competencies of Data Science. On the data engineering side, we began by choosing a topic that would add depth to our understanding of developing machine learning models in industries that were novel to us and outside of our domain expertise. We decided on the banking and financial sector since this industry is a hot topic in today's economic landscape across the board. Next, we reviewed Kaggle and datasets provided by our professor, Dr. Jacob Turner, that were open-source and robust in the data it provided. We settled upon the following dataset from UC Irvine's Machine Learning Repository, which matched the criteria needed to meet our project goals Bank Marketing. Finally, we preprocessed and cleaned our datasets, queried the data, built several machine learning models, and created an engaging and informative talking slide deck presentation using Canva, which you can view by clicking the link found below under the header "How to Interact with this Project."
Data. It's here. It's there. It's everywhere. The amount of data ready for an analysis is expansive, and advances in technology has heightened the levels of complexity and sohistication found in the information shared. Machine Learning Algorithms are gaining traction as powerful tools assisting bankers in determining the probability of success in financial strategy campaigns that help them retain their customers and increase profits from products and services they offer. The likelihood that one will require financial advice and products in their lifetime is inevitable. With this in mind, our group selected banking and finance as our topic of interest to explore how predictive analytics could be used to accurately determine the success of term deposit subscriptions in telemarketing campaigns. As progress in machine learning algorithms continues to develop and new technologies emerge, we can use various algorithmic tools and models to classify and interpret efficacy, validity, and percentage rates of improvements when needed. Predictive modeling working in tandem to accurately determine personalized financial products for specific consumers has innumerable advantages for companies, and has proven to be a wise investment for organizations looking to offer practical banking products to consumers who typically pay a visit to their local bank on rare occasions. After all, financial success doesn't happen haphazardly--it's the positive outcome of success that's a direct result of making data-driven decisions.
- R Studio
- Oldest Computer in the World: Our Brains
- Excel
- Click the following link Banking on a Call Canva Presentation to view our Canva Slide Deck Presentation discussing this topic.
Many thanks to our outstanding Professor, Dr. Jacob Turner, for supporting us through this year. His commitment to education and helping his students grow their knowledge in data science made a challenging subject to learn OK to be a challenge.
- Southern Methodist University Master of Science in Data Science
- Applied Statistics Professor: Dr. Jacob Turner