Marketing Sales Regression Analysis

Project Overview

Simple Linear Regression analysis for Darey.io assignment.
Goal: Recommend the marketing channel with the strongest ROI impact, explain model results clearly to a non-technical stakeholder`.

Key findings:

• R² = 0.999: Model explains 99.6% of sales variance
• TV has a high impact on Sales
• TV and sales met key model assunptions ( linearity, normality and homoscedasticity.
• Recommendation: Allocate more budget to TV for max ROI.

Assumptions

• Normality: Q-Q plot shows residuals follow the line, so normality assumption holds
• Homoscedasticity: Residuals vs Fitted plot shows constant variance around y=0

Files:

• Regression_analysis.ipynb - Jupyter notebook with EDA, model training, evaluation
• marketing.csv - Dataset with TV, Radio, Social Media, Sales columns

Environment Setup

1. Install Python 3.8+
2. Install dependencies:

   pip install pandas numpy matplotlib seaborn scikit-learn jupyter