Predicting Alumni Giving Rate using Linear regression

R Code | Powerpoint

Project Description

The project aims to identify factors influencing alumni donation rates using advanced statistical techniques. It developed a comprehensive linear regression model to predict alumni giving rates. This model would serve as a tool to understand the relationship between independent variables and alumni donations, thereby empowering universities to strategize and allocate resources more effectively.

Business Needs

Universities rely on alumni donations to support a range of initiatives, from scholarships to research programs. However, understanding the factors that drive higher giving rates is essential for optimizing fundraising efforts. By identifying these factors, universities can tailor their engagement strategies, fostering stronger relationships with alumni and encouraging higher levels of contribution.

Project Steps

1. Exploring Data: Initially, a thorough exploration of the dataset was conducted. The data summarized various variables, including graduation rates, student-faculty ratios, and % of classes under 20. Notably, the alumni giving rate exhibited a wide range, spanning from 7% to 67%.
2. Correlation Analysis: A comprehensive correlation analysis was further conducted to identify relationships between variables. The investigation revealed a positive correlation between graduation rates and alumni giving rates, while a negative correlation was observed with student-faculty ratios.
3. Model Building and Comparison: Further, the development of a multiple linear regression model was conducted. Model iterations were made, incorporating transformations of variables, such as squared and quadratic transformations. Models were rigorously compared based on metrics like root mean squared error (RMSE) to ascertain their predictive accuracy.
4. Final Model: The project's culmination resulted in the creation of the final linear regression model. This model incorporated quadratic transformations of relevant variables and demonstrated the highest prediction accuracy with the lowest RMSE.

Conclusion

The model's insights were translated into actionable recommendations. Notably, reducing student-faculty ratios and increasing graduation rates were identified as key strategies to enhance alumni giving rates. Practical steps, such as hiring more staff and expanding university resources will help achieve these goals.