Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/stefagnone/-air-france-sponsored-search-campaign-optimization

Optimized Air France's online advertising campaigns using advanced regression modeling and data analysis, uncovering actionable insights to maximize ROI and enhance ad performance.
https://github.com/stefagnone/-air-france-sponsored-search-campaign-optimization

advertising-analytics air-france data-science digital-marketing marketing-campaigns predictive-modeling python r regression-analysis roi-optimization scikit-learn sponsored-search

Last synced: about 1 month ago
JSON representation

Optimized Air France's online advertising campaigns using advanced regression modeling and data analysis, uncovering actionable insights to maximize ROI and enhance ad performance.

Awesome Lists containing this project

README 

        
## Project Overview

This project focuses on optimizing Air France's online advertising campaigns across Google, Yahoo!, MSN, and Kayak platforms. Using predictive modeling techniques and insightful analysis, the objective was to enhance ad performance, reduce costs, and maximize ROI.



The analysis included:

- Cleaning and exploring advertising data for trends and outliers.

- Developing and tuning multiple regression models to predict key campaign metrics.

- Identifying actionable recommendations for budget allocation and campaign adjustments.



This project demonstrates expertise in data-driven marketing, advanced modeling, and actionable business insights.



## Technologies Used

- **Python**: Data preprocessing, model training, and validation.

- **R**: Advanced statistical analysis and insights generation.

- **Pandas, NumPy**: Data manipulation and feature engineering.

- **Scikit-Learn**: Regression modeling and hyperparameter tuning.

- **Matplotlib, Seaborn**: Data visualization.



## Repository Structure

- `Code/`: 

  - Contains the Jupyter Notebook (`ASSIGNMENT_TEAM_9.ipynb`) with Python modeling and analysis.

  - R Markdown file (`Team9_Analysis in R-2.Rmd`) and its corresponding HTML output.

- `Data/`: 

  - `sample_submission.csv`: Initial data sample.

  - `ORIGINAL_VALUES.csv`: Cleaned dataset used for analysis.

  - `FINAL_SUBMISSION_ELASTIC_MODEL.csv`: Final predictions submitted to Kaggle.

  - `KEL319-PDF-ENG.pdf`: Original campaign data.

- `Documents/`:

  - `Team9_Analysis in R-1.html`: Insightful report with visualizations and recommendations.

 

## Key Insights

1. **Actionable Recommendations**:

   - Adjust keyword strategies to emphasize unbranded terms for broader audience reach.

   - Tailor campaigns based on specific platform advantages (e.g., demographic targeting on MSN).

   - Optimize bid strategies dynamically to maximize ROA for low-performing engines like Yahoo.

   

2. **Modeling Techniques**:

   - Comparative performance of Elastic Net, Ridge, and Decision Tree models for revenue prediction.

   - Feature engineering based on keyword categories and match types.

   - Hyperparameter tuning for maximizing predictive accuracy.



3. **Visualizations**:

   - Revenue trends across platforms.

   - Click-through and conversion rates for branded vs. unbranded keywords.

   - Platform-wise ROA comparisons to identify strengths and gaps.



These insights provide actionable recommendations for maximizing returns from sponsored search campaigns.



## Instructions

### Requirements

- Python 3.8+

- R 4.0+

- Required Python packages: `pandas`, `scikit-learn`, `matplotlib`

- R packages: `ggplot2`, `dplyr



## Contact

Connect with me on [LinkedIn](https://www.linkedin.com/in/stefano-compagnone98/) for more information or explore my other projects on GitHub.