1. Project Overview

• This project presents a comprehensive analysis of a hypothetical yet coherent dataset comprising 500 Google Ads campaigns.
• The refined dataset includes the following key columns:
  • CPC (Cost Per Click): Average cost incurred for each click.
  • Product Price: Selling price of the advertised products.
  • Quality Index: Quality score assigned to the ads.
  • Spend: Total amount spent on each campaign.
  • Conversion Rate (CVR): Percentage of clicks that resulted in conversions.
  • ROAS ≥ 3: Indicator for campaigns achieving a ROAS of 3 or more (Yes/No).

2. Data Preparation

Using Python, the dataset underwent several preprocessing steps, including:

• Data Type Transformation: Ensured that all columns were in the appropriate format for analysis.
• Data Cleaning: Removed null values, duplicates, and extra spaces to enhance data quality.
• Outlier Removal: Identified and eliminated outliers to minimize their impact on statistical analyses.

3. Statistical Analysis and Visualization

• A series of statistical analyses were performed to uncover relationships between variables. Visualizations were created to represent findings effectively, including:
  • Scatter Plots: To explore correlations between CPC and ROAS.
  • Histograms: To visualize the distribution of key metrics.
  • Box Plots: To identify variations and outliers across different campaigns.

4. Predictive Modeling

• A Logistic Regression classification algorithm was implemented as part of a machine learning approach to predict whether each campaign would achieve a ROAS of 3 or more (Yes/No). The model evaluation included:
  • Accuracy: Achieved an outstanding accuracy of 99.00%.
  • Confusion Matrix: Provided insights into true positives, true negatives, false positives, and false negatives.
  • ROC Curve: Illustrated the trade-off between sensitivity and specificity for the classification model.
  • Classification Report: Summarized precision, recall, and F1-score to assess model performance.

5. Conclusion

• This analysis underscores the importance of leveraging data-driven insights and machine learning techniques in digital marketing.
• By accurately predicting campaign performance using logistic regression, marketers can make informed decisions to optimize advertising strategies and maximize ROI.