Building Tree-Based Models: Practical Data Science Project
Extralearn is an initial stage startup that offers programs on cutting-edge technologies to students and professionals to help them upskill/reskill.
In this MIT PE project, I sucessfully build tree-based models that can assist ExtraLearn in predicting which leads are likely to convert to paying customers. I compared decision tree models and random forest models to get the best one. I endened up choosing as the best the tuned random forest model, which boasts an 85% Recall score along with well-balanced precision and F1 scores. This gives an strategic advantage to ExtraLearn that will allow the company to efficiently allocate resources while maintaining brand equity.
Furthermore, I complemented this classification model by conducting an Exploratory Data Analysis (EDA) at the outset of the analysis. I also provided conclussions and recommendations at the end to formulate effective marketing strategies based on our findings.
Tool: Visual Studio Code
Tech Language: Python
Visualization of the Tuned Decision Tree
