Topic 1: Exploratory Data Analysis automation

Description:

This project focuses on the critical role of pre-modeling analysis in MMM and explores ways to automate this phase for efficiency. Exploratory data analysis ensures that data is clean, relevant, and optimized for modeling, significantly impacting the quality and reliability of the final model results. This project aims to identify key pre-modeling steps, such as data exploration, transformation, and validation, and to automate them where possible.

Key attributes / Main competencies:

Experience with Spark & SQL is a plus

• Experience with common data science toolkits (R, Python, etc…) is a plus 
• Excel and PowerPoint skills required

Analytical and problem-solving skills

Learning Outcomes:

• Develop a structured approach to pre-modeling analysis in MMM
• Identify and automate key pre-modeling steps for improved efficiency and accuracy
• Improve modelling insights and recommendation

Topic 2: Improve Reporting Capabilities

Description:

This project focuses on enhancing the effectiveness and customization of reporting dashboards in MMM. The goal is to design a flexible dashboard architecture that can be reused and adapted to client-specific needs, reducing the time spent on data checks while maintaining data quality. Additionally, exploring the automation of PowerPoint presentation creation through Power BI will streamline the reporting process.

Key attributes / Main competencies:

• Dashboard architecture and design

• Power BI automation and presentation generation

• Data quality assurance and efficiency optimization

     Learning Outcomes:

• Develop a customizable, reusable dashboard for efficient data checking

• Automate the generation of presentations from MMM dashboards

Improve process efficiency without sacrificing data quality

Topic 3: Build MMM Models WITH Hierarchical Modelling

Description:

This project explores alternative model estimation techniques to improve model quality in MMM. It focuses on evaluating the impact of hierarchical modelling for regional segmentation. This investigation aims to determine how these techniques can improve the quality of the built models.

Key attributes / Main competencies:

• Understanding of data mining, machine learning, and statistical models including linear & logistic regression, PCA, clustering, etc…

• Experience with common data science toolkits (R, Python, etc…) is a plus 
• Excel and PowerPoint skills required

• Analytical and problem-solving skills

Learning Outcomes:

• Evaluate hierarchical modelling benefits for segmented markets

• Identify effective model selection criteria
• Balancing business/statistical relevance
• Model validation for bias and prediction power

• Suggest enhancements/improvements
  

Topic 4: Build MMM Models with Bayesian Regression 

Description:

This project explores alternative model estimation techniques to improve model quality in MMM. It focuses on assessing the benefits of the Bayesian regression approach to leverage priors and other domain knowledge obtained from experimentation or other analytics studies.

Key attributes / Main competencies:

• Understanding of data mining, machine learning, and statistical models including linear & logistic regression, PCA, clustering, etc…

• Experience with common data science toolkits (R, Python, etc…) is a plus 
• Excel and PowerPoint skills required

• Analytical and problem-solving skills

Learning Outcomes:

• Assess the effectiveness of time-variant coefficients and Bayesian methods

• Identify effective model selection criteria
• Balancing business/statistical relevance
• Model validation for bias and prediction power

• Suggest enhancements/improvements